(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "History of Color Photography"

History of 

PHOTOGRAPHY 




JOSEPHS. FRIEDMAN 



From the collection of the 



7 n 

m 



o Prelinger 

v JUibrary 

* p 



San Francisco, California 
2006 



HISTORY OF COLOR 
PHOTOGRAPHY 



By 
JOSEPH S. FRIEDMAN, Ph.D. 




1945 
THE AMERICAN PHOTOGRAPHIC PUBLISHING COMPANY 

BOSTON 



Copyright 1944 by 
American Photographic Publishing Co. 

Second printing, 1945 



Printed in the United States of America 
by The Plimpton Press, Norwood, Mass. 



PREFACE 

This book has been produced in response to an insistent demand from color 
workers for exhaustive information on the many forms of research that have 
developed the various color processes of photography into the usable condi- 
tion in which they are found today. This subject was covered very completely 
by Professor Wall up to the time of the publication of his famous book, "The 
History of Three-Color Photography," in 1925. But the demand for that 
work was so great that it has long been out of print and its information is no 
longer generally available. Even if it were, color photography has progressed 
so rapidly in the past twenty years that information as of that date could 
tell no more than half the story of today. 

To the stupendous task of ferreting out and compiling into coherent and 
usable form all this accumulated data, Dr. Friedman brings a splendid prepara- 
tion. After graduating from Harvard and taking his doctorate at the Uni- 
versity of Chicago, he plunged directly into color work on the staff of Tech- 
nicolor which was then evolving its famous process in Boston. Through the 
years he has been actively identified with the development of many forms of 
color photography, and is at present on the research staff of Ansco. He has 
long been known as a prolific and authoritative writer on this subject, and of 
late years his department in American Photography has been a general clearing- 
house of information about its latest aspects. 

This book will be found invaluable to anyone who needs the complete record 
of what has gone before in any existing department of color photography. 
Starting with the earliest ideas of colorimetry, it traces the development of 
all the laboratory and commercial processes by which color has been evolved 
to its present-day applications, enumerating the underlying principles, de- 
scribing the technique, and giving the history of the patents that have been 
issued concerning them. The record is as complete as it is humanly possible 
to achieve, and contains compactly compiled and correlated information that 
is nowhere else available without very extensive research. For anyone who 
wants to get a detailed and comprehensive picture of color photography as a 
whole, or who needs specific information about any of its special developments, 
no effort has been spared to make this book as complete and valuable as 
possible. 



CONTENTS 



chapter page 

i colorimetry i 

2 Subjective Color Reproduction 13 

3 Objective Color Reproduction 23 

4 Color Cameras 34 

5 Color Cameras, Type 3 41 

6 Color Cameras, Types 4 and 5 54 

7 The Photographic Emulsion 66 

8 Tripacks and Bipacks 82 

9 monopacks 94 

10 Kodachrome and Kodacolor 108 

11 Ansco Color and Agfacolor 125 

12 Screen Plates 135 

13 Formation of the Screen 147 

14 Processing Screen Plates 173 

15 Separation Negatives 197 

16 The Lenticular Process 214 

17 Lenticular Disclosures , 232 

18 Duplication of Lenticular Film 249 

19 Masking 273 

20 Chemical Toning 296 

21 Dye Toning 327 

22 Primary Color Development 354 

23 Coupling Color Development * 373 

24 Gasparcolor and Silver-Dye-Bleach Processes 405 

25 Carbon and Carbro 430 

26 The Transfer Processes 462 

27 Diazo Photography 487 

28 Bleach-Out Photography 494 
List of Periodicals and Abbreviations 505 
Index 507 



INTRODUCTION 

It is now almost twenty years since E. J. Wall published his monumental 
work "The History of Three-Color Photography." In that interval progress 
has been rapid. Viewpoints and emphasis have changed so much that the 
practice of 1944 uses techniques that received hardly any mention in 1925. 
Color photography passed from the use of the cumbersome one-shot camera 
with its delicate light-splitting device, to the monopack which could be ex- 
posed in an ordinary camera. It discarded the complicated positive processes 
which required precise registration of three separate images, and adopted the 
multi-layered material processed by color development or by silver-dye-bleach. 
It is somewhat ironic to recall that monopacks and color development were 
fully disclosed by 1913. And yet, while Mr. Wall was able to summarize the 
work done in this line up to 1925 in a few scattered paragraphs, it has since 
assumed such importance that these paragraphs are expanded to six chapters 
in the present text. 

The use of monopack film leaves the operator with a color reproduction of 
an original scene. The ultimate object, for the motion-picture industry at 
least, is the conversion of this into any number of duplicates, all of uniform 
quality. The duplication of a color reproduction presents new problems whose 
solutions are indicated in the chapter dealing with masking. This is a pro- 
cedure first introduced by Dr. Albert before the turn of the present century, 
for the photomechanical industry. Perhaps for that reason it was ignored 
by the photographer. When it was brought to his attention, he appraised it 
quickly as being "highbrow," and consequently scorned it. He made no 
serious effort until quite recently to study its possibilities. Consequently Wall, 
in his historical survey, found no reason to mention it by name, although he 
did describe procedures where masking was utilized. At this writing, there 
can be no further question of the usefulness of the technique. It complements 
the monopack film, for in combination with that material it seals the doom 
of the one-shot camera. 

The present text includes chapters on diazo and bleach-out photography. 
No great strides have been made in these fields to warrant the hope that they 
may some day offer a solution to the print problem. However they do suggest 
a possible extension of useful light-sensitive systems. The use of silver halides 
represents a constant challenge to the photographer and photographic chem- 
ist, for the silver halide system is extremely wasteful of light energy. When 
light is incident upon a photographic film, a large percentage of it is reflected 
off. Another substantial portion is transmitted. These represent sheer waste 
of light energy. Of the light that actually penetrates the emulsion layer, 



x INTRODUCTION 

only a small part is utilized for the formation of the latent image. The rest 
is lost by scatter. In a system sensitized by dichromate or by a fugitive dye, 
only the useful light is absorbed, the rest is transmitted freely. This could 
be utilized for a similar or other purpose. Attempts have been made to com- 
bine the high efficiency of the diazo and bleach-out systems with the intensifi- 
cation possibilities of the silver halides. These are discussed in the chapters 
on diazo and bleach-out processes. 

The history and development of color photography prior to 1925 has been 
very ably covered by Mr. Wall. Therefore in the present text it was deemed 
best to concentrate on the work done since that time. However each subject 
has been traced back, as far as possible, to its initial introduction. This 
was done mainly to show that the basic principles which underlie nearly all 
our present-day procedures, are now in the public domain. The keynote of 
present research is not so much a search for something new. It is rather a 
search to improve and make useful the suggestions of yesterday. True, we 
need men of genius to create new scientific marvels. But more than that, we 
need the dogged plodder who can work through discouragement and who 
can analyze failure. It is this man who will make practical the brilliant achieve- 
ment of the genius. 



CHAPTER 1 
COLORIMETRY 

V>OLOR photography can be identified with colorimetry, the science that 
deals with the specification and measurement of color. It is, in fact, a system 
of mechanical colorimetry, where the amount of silver developed in a pho- 
tographic emulsion, after it has been exposed to the light and color under 
consideration, is utilized as a measure of the intensity of a given primary 
present in that color. The fundamental ideas concerning colorimetry can 
therefore be of great importance in the practice of color photography. 

The common practice with regard to both colorimetry and photography, 
is to regard the subject from the strictly psychological point of view. This 
is due to the fact that, as sensed by man, color is a sensation and not a sub- 
stance. This fact was not recognized until very early in the nineteenth century, 
when Thomas Young read his famous paper to the Royal Society in 1802. 
The three-color theory bears his name because he was the first to emphasize 
this fact. Writing some forty years later, Maxwell had this to say, "It is 
almost a truism to say that color is a sensation, and yet by recognizing this 
simple fact Thomas Young realized clearly nearly forty years ago that the 
science of color must therefore be regarded as essentially a mental science." 
Stated broadly, the idea promulgated by'Young was that the sensation pro- 
duced by any one color could be matched by the proper mixture of three 
fundamental colors called primaries. Thus the infinite variety of hues present 
in the rainbow could each individually be matched by some mixture of three 
colors which are themselves also present in the rainbow. Young based his 
idea upon the hypothetical existence of three light-sensitive substances within 
the optical system of the human being. It is in this respect that he differed 
from his many predecessors who had actually experimented with trichromatic 
matching of colors. 

The idea of primaries, from whose mixture all other colors could be com- 
pounded, is a deep-rooted one. The ancient Greeks taught that all colors 
could be compounded by the proper blending of white, blue, red, and yellow. 
Aristotle, noting that the rainbow was produced when only gray clouds and 
sunlight were present, advanced the idea that there were but two primaries, 
white and black. This idea dominated the thinking in this field up to the 
seventeenth century. Antonius de Dominis as late as 1624 propounded a 
like theory. Maurolycus (1494-1577) distinguished four principal colors in 
the rainbow. These were reddish yellow, green, blue, and purple, each of 



2 HISTORY OF COLOR PHOTOGRAPHY 

which blended into the next by means of an intermediate color. This was a 
clear anticipation of Newton, who also distinguished seven principal hues. 
Leonardo da Vinci, the genius of the Reformation, considered red, yellow, 
green, blue, white, and black as simple colors, and all the others as com- 
pounded from these. The three-primary-colors scheme was advocated by 
Lambert (1728-1777), Tobias Mayer (about 1775), and Brewster (1781-1868). 
It is significant that these scientists all chose red, yellow, and blue as the 
primary colors. From our later knowledge, we can assume that in these cases 
the red primary was considerably pinkish in hue, while the blue was quite 
greenish. Even today, when we have a more thorough knowledge of the 
fundamentals, we still call the minus-green red, instead of magenta, and the 
minus-red blue instead of cyan. This is a relic of the days when the hues of 
the secondaries were so far from the theoretical that they could truly be called 
red and blue, rather than magenta and cyan. We may therefore infer that the 
seventeenth and eighteenth century scientists did not distinguish between 
the additive and the subtractive colors. The colors used in the additive proc- 
esses (red, green, blue) are termed primaries, while those used in the sub- 
tractive processes (cyan, magenta, yellow) are termed secondaries. The 
reason for this will be made apparent later. Most, if not all, the experiments 
with color matching were done by mixing pigments, and so those scientists 
who operated in this manner used secondary colors. This explains why an 
artist and scientist like Leonardo da Vinci, fell into this error. Those who 
matched colors by mixing colored lights, used the primary colors, since they 
operated with additive systems. Goethe in 1810, in his attack upon Newton, 
also fell into the same error of confounding secondary with primary colors. 

Probably the first person who carried out extensive work upon trichrome 
color reproduction was Tobias Mayer. He left a notebook with copious notes 
in which he tried to determine the approximate number of distinguishable 
colors that could be compounded from three pigments. As stated above, he 
used red, blue, and yellow for his basic colors, but from the nature of his 
experimental work, it seems fairly obvious that his red and blue were really 
the red and blue of the printer, that is a pinkish red, and a greenish blue. 
His color chart represented each different hue by a formula of the type r a/ /b y , 
and for the most part a + /3 + y equaled 12. The first person to put these 
ideas into practice was the printer J. C. LeBlon, who in 1722 used red, yellow, 
and blue printing plates to make colored reproductions. But none of these 
gentlemen realized that the matching of colors was purely a psychological 
process, that the fusion of the three primaries took place within the optic 
system of the human, and that the sensation of color was purely a figment 
of the imagination. This was Young's contribution, and for this reason the 
three-color theory rightly bears his name. 

The Young theory was extended considerably by Helmholtz in Germany 
and by Maxwell in England. Maxwell was the first to demonstrate that the 
then newly developed art of photography could be utilized to give a graphic 



COLORIMETRY 3 

and automatic method for the determination of the relative amounts of the 
various primaries that were present in any given color. In two lectures which 
he gave in 1857 and 1859 he demonstrated the photographic procedure. In 
his experimental work he used a blue filter made by interposing an ammoniacal 
solution of copper sulphate, a green filter which consisted of a solution of 
copper chloride, a yellow filter, and a red filter that was composed of iron 
sulpho-cyanide. Thus his primaries were the correct ones, although it is not 
apparent just why he used a fourth separation made behind a yellow filter. 
In this practice, however, he anticipated the four-color printers of modern 
times by fully fifty years. From his four negatives he made four positives, 
each of which he projected in superposition upon a white screen through the 
same filters that he used in making the negatives. The synthesis process was 
additive. 

To Maxwell, the preparation of a colored photograph was but incidental to 
the study of colorimetry, in which he was keenly interested. He was quick 
to discover that colorimetry was identifiable with color photography, and that 
this last was merely applied colorimetry. But the state of the photographic 
art at that time was such that its practical application was a tedious job. 
There were no one-shot cameras available, so that the different "separations" 
had to be made one after the other. Collodion emulsions were the only ones 
which were known. The procedure, therefore, was equivalent to the wet- 
plate processes, hence exceedingly slow. The exposure through the blue filter 
was six seconds; through the red it was eight; and through the green it was 
twelve seconds after the filter had been considerably diluted. It is no wonder 
that Maxwell was satisfied to treat the entire affair merely as an experimental 
confirmation of his work on colorimetry. 

Since this is so closely allied to color reproduction, it is well to review some 
basic facts. It has its basis in the. Young theory of color sensation, in that it 
stipulates that all color sensations can be matched by a suitable mixture of 
three widely differentiable colors. Young and his successors postulated that 
within the optic system of the human there were three receptors each of which 
was predominantly sensitive to the light of a single primary. Newton's 
discovery of the dispersion of light into a spectrum more thoroughly convinced 
the members of this school. In accordance with the Newtonian experiment, 
white light when passed obliquely into one side of a glass prism emerged from 
the other side, not as a beam of white light, but as a band of colors. This 
band was always constituted in the same manner. One end of it was colored 
blue-violet, while the opposite end was colored red, or rather an orange-red. 
Between these two extremes the colors all blended into one another, so that 
the blue-violet became blue, then blue-green, and this in turn became a green. 
The green shaded off into a yellow-green, then yellow, orange, and finally 
turned red. Each color differed in hue by an infinitesimal amount from the. 
color that was situated immediately adjacent to it on the right and on the 
left. It may not be possible to distinguish two adjacent colors, so close is 



4 HISTORY OF COLOR PHOTOGRAPHY 

their match; but it is an easy matter to distinguish the color on the right of 
a given one from the color on its left, especially if the middle color remains 
between the two. The wave theory of light has made it very easy to differen- 
tiate the colors from one another. The wavelengths of each hue differ from 
each other, and these values can be determined to an accuracy of one part 
in ten million. Thus the wavelength of the D 2 line of sodium is 5889.977, 
while the wavelength for the A line is 5895.944. These values are in Angstrom 
units, which are equal to 0.1 m/x. The eye is sensitive to colors that lie in 
the range between 400 and 700 m/x. Thus there are an infinity of colors and 
hues which the eye can distinguish. Of course, if the eye is accosted by two 
colors whose wavelengths differ but slightly (and as far as the eye is concerned 
a difference of 5 m/x is very slight), it would be extremely difficult for it to 
differentiate between the two. As a matter of fact Newton divided the band 
into only seven distinct colors. But the fact remains that theoretically a 
difference in hue does exist between two colors whose wavelengths differ by 
as little as one mju. 

Because physics demands the existence of this infinite range of different 
colors, color specialists have insisted upon a three-color theory. The eye 
cannot have present at every point of its sensitive area an infinity of light- 
sensitive elements each of which would be sensitive to light of but a single 
wavelength. But the eye does distinguish between hues, and experiment has 
shown that a simple mathematical equation connects any four distinct colors, 
the only reservation being that no one of these should be matchable by a 
mixture of two of the others. The equation is as follows: 

mA + nB + pC + qD = o 

where m> n, p, and q are constants that could be positive, zero, or negative 
(but not all of them could be equal to zero at the same time), and A y B, C, 
and D are the four colors. This means that it is possible to match any one 
color by some combination of the other three. This is the real and experimental 
basis of the trichrome theory of colorimetry. It is independent of any hy- 
pothesis which attempts to explain why. Experimentally, the procedure can 
be carried out in this manner. One unit of the hue A is passed into one half 
of the photometric field of a color-matching instrument, while into the other 
half there are passed n/m units of the color B, p/m units of the color C, and 
q/m units of the color D. If any of the coefficients in the equation are negative, 
then these colors are added to the color A and the mixture is matched by a 
mixture of the remaining primaries. 

When two colors are mixed together, the colors being in the form of light, 
a third color results. The eye cannot distinguish in the third color either of 
the originals. In this respect the eye differs from all physical color-analyzing 
apparatus, in which the light is dispersed into spectra. These will always 
consist of the sum of the spectra of the component lights. The sum of two 
color sensations is also a color sensation, but it is not the linear sum of the 



COLORIMETRY 5 

other two sensations. Psychology does not have a conservation law of the 
type that plays so important a role in physics. 

The mixture of two colors is a third color that is different from either of 
the other two, and in which it is not possible to detect them. In particular, 
it is not possible for the eye to tell whether a given color is the sensation pro- 
duced by monochromatic light (light of a single wavelength), or whether it 
is produced by the light of a whole range of colors. For instance the color 
of the A filter is well known. It is red. When the light that this filter trans- 
mits is sent into a spectroscope it is seen that it consists of a mixture of all 
the colors whose wavelengths are greater than 600 m/x. The red due to the 
A filter is therefore a complex mixture containing an infinite number of hues. 
If one half of the photometric field of a color-matching apparatus is illuminated 
by monochromatic light of wavelength 610 m/x, and the other half by the 
light transmitted by the A filter, the two will match perfectly. The eye will 
not be able to distinguish between them. In color-matching procedures it is 
immaterial therefore, whether monochromatic light or color blends are used 
as the primaries. The same results will be obtained in both cases. 

Not all the colors to which the eye is sensitive appear in the spectrum of 
white light. In particular, the mixtures of the reds and the blues are not 
present in the spectrum. These are the pinks and the violets. As a matter 
of fact, the mixture of two spectral colors gives rise to colors that can be 
matched in hue only by some color in the spectrum (if that mixture does not 
consist of red and blue). This brings us to a discussion of the other properties 
of color. Suppose we are given a large number of colored patches to classify. 
The first and most apparent classification is that of hue. All the reds can be 
grouped together, then all the greens, etc. But it is also possible to classify 
them in another manner. It is possible to compare a red with a green and 
say that the intensity of red in the red is equal to the intensity of green in 
the green. By this classification we recognize that colors can have different 
degrees of intensity or brightness. It is also possible to consider the reds 
as a class and to compare two reds with identical hues, but which differ from 
each other in brilliance. Thus we may consider that one red appears to be 
quite pure, while another appears to be paler, as if it were a mixture of red 
and white. This type of classification is in terms of saturation. It is really 
quite significant that psychologists have arrived at a three- variable description 
of color, and at a trichrome analysis of color. Most probably the search for 
three variables was greatly influenced by the fact that three primaries were 
sufficient for color matching. Colorimetry attempts to give quantitative values 
to the three variables. 

If monochromatic light is passed into one half of the photometric field and 
a suitable mixture of three independent colors (in that no one of them can be 
matched by a combination of the other two) is passed into the other half, it 
will be possible to match every monochromatic color in the spectrum. When 
this is done, it will be possible to draw a graph whereby relative intensities are 



6 HISTORY OF COLOR PHOTOGRAPHY 

plotted against the wavelength for each primary. This will give rise to three 
curves, which can be considered to be measures of the stimuli produced in 
the three postulated sensitive elements in the optic system. Such curves are 
called stimulus curves. Maxwell was the first to carry out such an experiment. 
He was followed by a host of others, notably by Konig and Dieterichi, Abney, 
Wright, Guild, etc. The results showed a remarkable similarity. The Inter- 
national Committee on Illumination has adopted a set of such curves as a 
standard. It may appear that choosing one set of such values as a standard 
is rather arbitrary, but that is not so. It is true that the specific values for 
the coefficients depend upon the choice of the primaries, but then it is possible 
to determine these values for one set of primaries in terms of another. It 
will be found that this same relationship holds for any other color that may 
be matched in terms of the two sets. Knowing the relationship between two 
sets of primaries makes the values determined by one set convertible into 
the other. The I.C.I, values have a definite significance in that one of the 
primaries has been chosen so that it corresponds exactly to the sensitivity 
of the normal average eye. This is termed the Y co-ordinate, and its value 
can be used directly as a quantitative measure of the brightness of any color. 
Thus one of the variables of color specification has been given a numerical 
value. The I.C.I, tristimulus values of the spectrum colors are given in the 
table below, to a degree of accuracy that allows their use in calculations for 
our purposes. 

Wave- v rr r 



400 


0.014 


0.000 


0.068 


410 


0.044 


0.00 1 


0.207 


420 


0.134 


0.004 


0.647 


430 


0.284 


O.OI2 


1.386 


440 


0.343 


0.023 


1.747 


450 


0.336 


0.038 


1.772 


460 


0.291 


0.060 


1.669 


470 


0.195 


0.091 


1.288 


480 


0.096 


0.139 


0.813 


490 


0.032 


0.208 


0.46s 


500 


0.005 


0.323 


0.272 


510 » 


0.009 


0.503 


0.158 


520 


0.063 


0.710 


0.078 


530 


0.166 


0.862 


0.042 


540 


0.290 


0.954 


0.020 



Wave- 
length 


X 


F 


Z 


55o 


o.433 


o.995 


0.009 


560 


o.595 


o.995 


0.004 


57o 


0.762 


0.952 


0.002 


580 


0.916 


0.870 


0.002 


590 


1.026 


o.757 


O.OOI 


600 


1.062 


0.631 


O.OOI 


610 


1.003 


0.503 


0.000 


620 


0.854 


0.381 


0.000 


630 


0.642 


0.265 


0.000 


640 


0.448 


o.i75 


0.000 


650 


0.284 


0.107 


0.000 


660 


0.165 


0.061 


0.000 


670 


0.087 


0.032 


0.000 


680 


0.047 


0.017 


0.000 


690 


0.023 


0.008 


0.000 


700 


O.OII 


0.004 


0.000 



These values may be interpreted as indicating the relative response which 
each of the three sensitive elements in the optic system gives when stimulated 
by monochromatic light. 

The primaries used in any color-matching experiment are not restricted 
except that they must be independent of each other. It is therefore permissible 
to make a change of co-ordinates provided a linear relationship holds between 
the new and the old. This we proceed to do in the following manner. We 



COLORIMETRY 7 

define the co-ordinates x, y, and z by the equations x = X/R, y = Y/R, and 
z = Z/R, where R- X+Y + Z. A simple calculation will demonstrate that 
x+y+z= 1, so that only two of these are really independent. It is possible, 
therefore, to use two of the three co-ordinates as new variables, and for the 
third variable keep the Y component, since that is a numerical designation 
of the brightness. It is customary to choose x, and y. These can now be 
plotted as rectangular co-ordinates. From the table, the value of R for the 
wavelength 400 would be 0.014 + 0.000 + 0.068, or 0.082. Then x will be 
given by 0.014/0.082, and y by 0.000/0.082. Thus x will have a value of 
0.17 and y a value of 0.00. This will give rise to one point on the diagram. 
In a similar manner the wavelength 550 will give rise to a value of 0.433 
+ 0.995 + 0.009 or 1.439 * or &> anc * x wm * be equal to 0.433/1.439 or 0.30, 
while y will be equal to 0.995/1.439 or 0.69. This will give another point on 
the diagram. The values for x and y for the entire spectrum are tabulated 
below. 



Wave- 






Wave- 






Wave- 






length 


X 


y 


length 


X 


y 


length 


X 


y 


400 


0.17 


0.00 


500 


O.OI 


0.54 


600 


0.63 


0.37 


410 


0.17 


O.OI 


5io 


O.OI 


0.75 


610 


0.67 


0.33 


420 


0.17 


0.0 1 


520 


0.07 


0.83 


620 


0.69 


0.31 


430 


0.17 


O.OI 


530 


0.16 


0.80 


630 


0.71 


0.29 


440 


0.16 


O.OI 


540 


0.23 


0.7s 


640 


0.72 


0.28 


45o 


0.16 


0.02 


5So 


0.30 


0.69 


650 


o.73 


0.27 


460 


0.14 


0.03 


560 


0.37 


0.62 


660 


o.73 


0.27 


470 


0.12 


0.06 


570 


0.44 


0.56 


670 


o.73 


0.27 


480 


0.09 


0.13 


580 


0.51 


0.49 


680 


o.73 


0.27 


490 


0.05 


0.30 


S90 


0.58 


0.42 


690 
700 


o.73 
0.74 


0.27 
0.26 



It is to be noticed that at the two ends of the spectrum the points crowd 
each other severely. Had we carried the calculations out to the 4th or 5 th 
decimal point, slight differences would be seen to exist between all the co- 
ordinates, but for our purposes it is not necessary to carry the calculations 
to any greater accuracy. The graph of the points due to the spectral colors 
is shown in Fig. 1. The point C in the diagram represents the locus of the 
light adopted by the International Illumination Committee as a standard 
white approaching daylight. A diagram such as is shown in Fig. 1 is called 
a chromacity diagram, for it indicates the dominant hue and the saturation 
of any given color. Pure white, which is a theoretical mixture of equal parts 
of the three primaries, would be located at the point x = y = z = 0.33. This 
is not far away from the locus of the point C. 

A chromacity diagram is extremely useful in that it can graphically demon- 
strate all the known facts concerning additive color mixing. The points on 
the curve that depicts the spectral colors denote complete saturation or purity, 
and these are given a value of 100 per cent. Not all colors will be located upon 
this curve. In fact, the great majority of them will have positions within the 
area included in the curve. Suppose that the point A in Fig. 1 denotes a 



8 



HISTORY OF COLOR PHOTOGRAPHY 



color whose co-ordinates are (0.23, 0.39). To obtain the dominant hue of 
this color, a line is drawn from the point C, which represents the standard 
white, to the point A, and this line is continued until it intersects the spectral 
curve. The point of intersection P corresponds to a color whose wavelength 
is 506 mju. This will be the dominant hue of the color whose locus is at A. 
To find the purity of the color A y we determine the ratio of the line AC to CP, 
roughly 2/9. The purity of the color is therefore 22 per cent. This color 




*=*/r 



can be matched identically by a mixture of seven parts of white light (day- 
light) with two parts of monochromatic light of wavelength 506 m/i. This is 
because the line AP is 7/9 CP. 

Let us see now the effect of mixing two colors, R and S, in the ratio of three 
parts of R to one part of S. We first of all plot the two points upon the chro- 
macity diagram. The next step is to connect the two points by a straight 
line RS and divide this line into 3 + 1 parts. The result of mixing R with S 
will be a color whose locus lies somewhere upon the line RS y say at a point T. 
This point will divide the line RS so that ST will be equal to three times TR. 
That is, the new color will be situated three-quarters of the distance toward 
R 7 the dominating color. It is seen that the line SR passes through the point C. 
The segment SC is roughly equal to the segment CR } so that white light of 
color C will result when equal parts of lights R and S are mixed. Thus the 
result of mixing two colors can under the proper conditions give rise to white. 



COLORIMETRY 9 

Such colors are called complementary. It is readily seen that there are infinite 
numbers of such pairs of complementary colors. The line SCTR intersects 
the spectral curve in two places, one at a point which corresponds approxi- 
mately to 586 m/x, the other at approximately 484 m/x. If equal quantities 
of monochromatic lights of these wavelengths are mixed, the result will be 
the sensation of white, despite the fact that from a physical point of view 
white light would contain all the wavelengths. In this manner it is possible 
to determine all the pairs of complementary colors. Thus the point 700 m/x 
will lie on one end of the line through C while at the other end will lie the 
point 492. Since the two segments are not equal to each other, it will require 
more of the light of wavelength 492 than of wavelength '700 in order to obtain 
a white. In a similar manner the wavelength 400 m/i has for its complement 
a color whose wavelength is 567. 

It is seen that in the diagram in Fig. 1 the point 400 has been connected 
with the point 700 by a straight line. This line represents the locus of those 
colors that can be formed by a mixture of blue (400 m/x) with red (700 m/x).. 
These are the pinks and violets which the eye can see as distinct colors, but 
which have no counterpart in the spectrum. There are no spectral matches 
to the colors which lie along this line, hence there can be no spectral comple- 
ments to the colors in the range from 492 to 567 m/x. The color 492 is a green- 
ish blue, while that of 567 is a pure lemon yellow. It is not possible to designate 
colors along this line by wavelengths, so it is customary to denote them by 
the wavelength of the complementary colors and place a small c after the 
number. Thus the color Q, which is complementary to the color P would 
carry the designation 506c, meaning that it is a color that is complementary to 
the color 506 m/x. 

We have stated above that the eye cannot tell whether a given color is 
monochromatic or whether it is composed of a mixture of a great many colors. 
This is made manifest when one plots the location of the hue transmitted by 
the A filter. The method of calculation is very interesting so it will be given 
here to serve as an example. We will consider that we are using daylight as 
our standard white. This is a color whose spectral distribution is well known. 
The intensity of each wavelength as 'it is present in this light is multiplied by 
the tristimulus value for that wavelength. This gives a measure of the stimu- 
lus intensity that each wavelength can give. But the A filter passes only a 
definite percentage of the light, absorbing the rest. The product above must 
therefore be multiplied by the percentage transmission of that wavelength 
by the filter in order to obtain the tristimulus value due to the particular wave- 
length passed by the filter. In this way there can be determined the values 
of the X, Y, and Z coefficients. If these values be determined for every wave- 
length that the filter transmits, and all the X's added together, there will be 
determined the value of the X coefficient for the entire transmission of the 
A filter. The F, and the Z terms can be determined in a like manner. From 
these, the values of x and y can be calculated. 



io HISTORY OF COLOR PHOTOGRAPHY 

The transmission of the A filter begins at 570 mju, where a total of 0.25 per 
cent of the light is transmitted. Light of this color has a relative intensity 
of 107.18 in daylight. The tristimulus values for this wavelength are X = 0.762, 
Y = 0.952, Z = 0.002. Hence the contribution of this wavelength to the 
total X value would be 0.0025 X 107.18 X 0.762 or 0.20. The Y contribution 
would be 0.03, and the Z contribution would be completely negligible. In 
the table below is listed the rest of the values. 



Wave- 
length 


X 


Y 


Z 


Wave- 
length 


X 


Y 


Z 


570 


0.20 


0.03 


0.00 


640 


61. 1 


24.1 


0.00 


580 


iS-« 


15.1 


0.0 


650 


40.9 


14.8 


0.0 


S90 


62.5 


46.1 


0.1 


660 


24.4 


9.6 


0.00 


600 


94.5 


55-9 


0.1 


670 


13.7 


5-7 


0.0 


610 


109.6 


55-2 


0.1 


680 


7.6 


2.8 


0.0 


620 


102. 1 


45-7 


0.0 


690 


3-8 


1.4 


0.0 


630 


81.5 


33.8 


0.0 


700 


2.0 


0.7 


0.00 



Totals 619.5 310.0 negligible 

The sum of X, F, and Z, which is R, is 930. Therefore x is equal to 620/930 
or 0.67, while y is equal to 310/930 or 0.33. This is the point A in Fig. 2. 
It is identically equivalent to monochromatic light of wavelength 610 m/x. 
A similar calculation for the B and the C5 filters gives the following values: 

B filter x is equal to 0.28 y is equal to 0.66 
C5 filter x is equal to 0.14 y is equal to 0.07 

These are the points B and C5 in Fig. 2. A glance at the diagram indicates 
that using daylight as our source, the A filter can be identified with mono- 
chromatic light of wavelength 610 m/x, the B filter can be matched by mono- 
chromatic light of wavelength 546 mju mixed with approximately 15 per cent 
daylight, while the C5 filter can be matched by a mixture of one part of day- 
light and nineteen parts of light with a wavelength of 469 mju. 

Consider now a colorimeter where these are the primaries used. One half 
of the photometric field is illuminated by light of the unknown composition, 
and the other half is illuminated by a mixture of these primaries, the ratios 
being changed until a match is obtained. The question now arises how ac- 
curately will such combinations measure all the colors known to man. Of 
course it must be understood that the source of light behind the niters is day- 
light. Only those colors can be matched whose loci are located inside of the 
triangle ABC$. Any color which is located outside the confines of this triangle 
cannot be matched. A study of the triangle indicates that very accurate 
rendition will be obtained of all the colors, including the spectral, whose 
dominant wavelengths are greater than 540. Below that wavelength there 
is a very sharp decline in the accuracy of matching the purity of the colors. 
Thus a color with a dominant wavelength of 520 m/z will never be matched 
in a saturation or purity greater than 40 per cent. This continues until about 



COLORIMETRY 



ii 



the wavelength 485 mju, when the best purity obtainable becomes 50 per cent. 
At 480 it is almost 70 per cent, and at 470 it is again at a maximum at about 
90 per cent. Beyond this the efficiency falls off again. It is impossible to 
obtain accurate rendition of the pinks and violets, as far as purity is con- 
cerned, although the efficiency of reproduction never falls below 60 per cent. 
Very fortunately for us, colors as we are accustomed to experience them in 
nature, are never spectral in quality, so that it is very seldom necessary for 



.9 










^szo 




.8 




>*S30 






iSJO 


N^O g 




^ .7 




A*£ S0 




II , 




Jvvw 




% 


>50Q 


h V 


sio 


2> .5 




/ \ N 

/ 1 


s\$"SO 


.« 




/ I 








/ ?° 


Am^io 


.3 


■ t490 


J^^\^>%o 






I / 

1 / ^ 


^^"^ ^6700. 


.2 




I / ^ — 




./ 










<<5> 




x- x /r 



.1 .Z .3 M--.5 .6 .7 

FIG. 2 

* 

us to attempt to duplicate colors outside of the triangle. The results in actual 
practice are therefore far better than would be expected merely from a glance 
at the triangle. 

In the table below are listed the dominant wavelength and the purities of 
some Wratten filters, all of which have, at one time or another been recom- 
mended for separation work in color photography. 



A. Red filters: 



No. 25 

29 
70 



Dominant Hue 
610 
631 
628 



Purity 

100% 

100 

100 



12 HISTORY OF COLOR PHOTOGRAPHY 

B. Green filters: 







Dominant Hue 


Purity 




58 


546 


85 




61 


543 


90 




62 


536 


98 




74 


535 


97 


C. Blue filters: 










47 


469 


95 




49 


466 


99 




So 


566c 


92 



These filters are plotted on a chromacity diagram in Fig. 3. 
•9 




0.3 04 

FIG. 3 



CHAPTER 2 
SUBJECTIVE COLOR REPRODUCTION 



I 



N the previous chapter it was pointed out that any color whatsoever can 
be matched by the suitable mixture of three basic colors. This is an experi- 
mental fact that is not concerned with any possible hypothesis used to ex- 
plain it. Young was the first to propound that the explanation lay in the 
human brain and that the optic system contained three sensitive elements 
which were differentially affected by the three primaries. In this respect 
Young differed from his many predecessors. Young's ideas were further de- 
veloped by Helmholtz and Maxwell, the latter being the first to put into actual 
practice the consequence of the idea. But it is extremely interesting to note 
that the experimental procedure adopted by Maxwell was not a true applica- 
tion of the Young hypothesis. He failed to recognize that the sensation curves 
which he obtained in the course of his work on colorimetry, were the basis 
for the theory of three-color reproduction. It was F. E. Ives who recognized 
this, and who incorporated it into his patent of 1890 (U.S.P. 432,530). The 
curves Ives reproduced in his patent specifications were the color-mixture 
curves of Maxwell, who used an arbitrary set of primaries. Of course it must 
be remembered that there is no one set of basic colors, but an infinity of 
sets of such, each of which is convertible into the other. The Ives disclosure 
let loose a rather lengthy discussion from which very little seems to have 
resulted. The poor results that were obtained when using filters for color 
separation whose transmissions corresponded to the color-mixture curves, 
soon forced the subjective idea of color reproduction into the discard, and the 
further development of the subject of color photography appears to have 
proceeded along objective lines. But although the practical work in this 
field shifted away from the original Young theory, the interest in it continued. 
The literature is fairly replete with articles upon the theory of three-color 
reproduction. Especially, there has been a series of articles which re-opened 
the entire problem, and which bids fair to give the subjective processes a real 
test. Typical of these is the article by A. C. Hardy and F. L. Wurzburg, Jr., 
which was published in the Journal of the Optical Society of America, Volume 27 
(1937), page 227. 

In this study, use is made of the tristimulus values discussed in the last 
chapter. Let X r , F r , Z r , be the tristimulus values for unit quantity of the 
red primary; let X g , Y 0} Z g , be the values for the green primary; and let 
X by F 6 , Z 6 , be the values for the blue primary. Let the mixture of the 
three primaries at some point of the picture, be r units of the red, g units of 

*3 



14 HISTORY OF COLOR PHOTOGRAPHY 

the green, and b units of the blue. If X', Y', and Z 1 be the tristimulus values 
for the color depicted at this point, then we can write the following equation: 

(l). X' = rX r + gX r + bX r 

Y' = rY + gY + bY g 
Z' = rZ b + gZ b + bZ b 

Let the spectral distribution of the light that is reflected from the correspond- 
ing point of the subject be E, so that the tristimulus values at this point on 
the subject will be given by the form: 

(2). X = Cmik 

Jo 
Y = fcvstk 

Z = f*Ezd\ 

where x, y, and z are the distribution functions of the observer. Perfect color 
reproduction will be obtained when X is identical with X' y Y with F', and 
Z with Z'. In practice there is allowable a certain tolerance, so that it becomes 
possible to write as a condition for perfect reproduction 

(3). X' = kX, Y' = kY, Z 1 = kZ 

This must be true regardless of the form of the function E. 

The first step is the preparation of separation negatives. Let the effective 
spectral sensitivity of the three photographic materials be 5 r , S g , and Sb. 
These values are defined as the products of the spectral sensitivity of the three 
emulsions, by the transmissions of the filters. Let the exposures impressed 
on the three emulsions when acted upon by E be 2J r , X and 2&, where 



(4). S r = CESrdX 
Jo 

Jo 
2 fe = f^ESbdK 



Without making any assumptions regarding the subsequent procedure, let 
the process be one in which r, the amount of the red primary in the reproduc- 
tion, depends only upon S r and is independent of 2^ and 2&. This can be 
done if the product of the contrast scales of the positive and negative proc- 
essing be always equal to one, and if the processing is always maintained 
upon the straight-line portion of the H & D curve. Therefore it follows that 



(5). r = krZr = k r f" ESjX 

Jo 

g — kgSfg = kg I ESgdX 

b = j& 6 2& = h rEStdX 
Jo 



SUBJECTIVE COLOR REPRODUCTION 15 

Substituting these values of r, g, and b into equation (1), and by (3), we get 
(6). k r X r f™ESrd\ + k X g £ES d\ + hXtCESjtk = f™Exd\. 

with similar expressions for the other two terms. If all the colors are to be 
reproduced faithfully, equation (6) must be valid regardless of E. This will 
be true if at every wavelength 

(7). k r XpSr + kgXgSg + kfiXhSb = X 

krYrSr « i g YgS g + hYiSt = y 

kfZfSr + kgZgSg + ^&Z&5& = 2 

In these equations X{, Yi, and Zi are the tristimulus values for unit amounts 
of the primaries, S r , S , and Sb are the spectral sensitivities of the emulsions 
used, and x y y, and i are the distribution functions of the observer in the 
special colorimetric system that he is using. The values for the constants k 
can be determined by experiment by assigning to them values which will 
make a fit for one color, and all other colors will then fall in line. 

An inspection of these equations shows that if the tristimulus values of any 
primary were multiplied by a constant, it would merely alter the value of the 
constant associated with the primary. These equations may therefore be 
rewritten in terms of the trichromatic coefficients of the primaries. Since 
it is the forms of the S functions that are of interest, rather than their absolute 
values, the constants may be discarded. In that case the fundamental condi- 
tions for exact color reproduction become 

(8). XtSr + XgSi+XA-x 
YrSr+YgSg + YtSb-y 

ZrS r + ZgSg + ZbSb = Z 

In the additive processes the primaries are easily identified, and their trichro- 
matic coefficients remain constant throughout the reproduction. Each primary 
can be projected individually and the tristimulus values measured with, a 
colorimeter. In accordance with I.C.I, procedure, this can be done by a 
determination of the energy distribution of the reflected light, and this is done 
by a determination, at each wavelength, of the product of the energy emitted 
by the lamp, the transmittance of the filter, and the reflectance of the screen. 
From this the trichromatic coefficients can be calculated in the manner used 
in the last chapter, to determine the characteristics of the colors transmitted 
by the filters. 

It is possible to solve equations (8) for S r , S oy and Sb, thus 

(9). S r - (YgZb - YbZ )x + (X b Zg - X g Z b )y + (XgY b - X b Yg)Z 
S = (YbZ r - Y T Z h )x + (X r Zb - XbZ r )y + (X b Yr - X r Yb)z 

S b = (Y r Zg - YgZ r )X + (XgZ r - X r Zg)y + (X r Yg - XgY r )Z 

In these equations we can insert the values for X r , F r , Z r , etc., after we have 
chosen the set of analysis primaries. 



16 HISTORY OF COLOR PHOTOGRAPHY 

In a reproduction system, the primaries should be chosen so that the area 
included within the triangle whose apices are the loci of the colors in the 
chromacity diagram, should be as large as possible. One good set of such 
primaries would be three monochromatic colors of wavelengths 700 m/j, 
535 mju, and 400 nogu. Selecting these for our set, we then have these values 
for the constants: 

700 m/i: X r — 0.74, Y T = 0.26, Z r = 0.00 
535 m/x: X = 0.19, Y = 0.78, Z g = 0.03 
400 mju: X b = 0.17, Y b = 0.01, Z b = 0.82 

Inserting these values into equations (9) there results 

(10). S r = o.6i# — 0.153; — 0.132 
S = — 0.21 x + 0.61^ + 0.13s 
Sb = 0.01 x — 0.02^ + 0.532 

To solve these equations we insert the values for x, y, and z. These are the 
relative stimulations that result when unit amounts of monochromatic light 
reach the eye. The result for the average eye has been adopted as a standard 
by the International Commission on Illumination. From the table on page 17 
it is seen that at the wavelength 400 mji, the values for x, y, and z are 0.01, 
0.00, and 0.07. Substituting these in (10) we get S r = 0.00, Sb = 0.03, and 
S g = 0.01. For the wavelength 500 mju, x, y, and z have the values 0.00, 
0.32, and 0.27 respectively. For this wavelength S r = — 0.08, Sb = 0.14, and 

Sg = 0.23. 

If the sensitivity of the emulsion is uniform for the entire range from 400 
to 700 mju, then these curves (Figs. 4, 5, and 6) represent the relative trans- 
missions of the taking niters. Unfortunately the sensitivity of the panchro- 
matic emulsions does not reach this ideal state. And equally unfortunately, 
the spectral sensitivities of the different emulsions are matters of information 
that the manufacturer does not consider to be worthy of the attention of the 
photographic technician, for the curves are not made public except in the 
form of very general wedge spectrographs. It is not possible therefore to 
calculate the transmissions for a set of ideal filters that will yield final sensi- 
tivity curves approaching those of S r , S gy and Sb- The product of the intensity 
of a given wavelength present in the light source, with the sensitivity of the 
emulsion, and the transmission of the filter should give a value for that wave- 
length equal to the value depicted in the sensitivity curves. 

It is to be noticed that the curve for S r starts off with a value of zero at 
400 m/x, then takes on a negative value, reaching a low of minus 0.09 at 510, 
beyond which it begins to rise until it reaches a value of zero at 535 m/i. 
From this point on it continues to rise to a maximum of plus 0.62 at 610, 
beyond which it begins to fall off again to zero at 700 mju. It is impossible 
at the present time to prepare emulsions that have negative sensitivities in a 
given range. Perhaps when we learn more concerning desensitization, we 



SUBJECTIVE COLOR REPRODUCTION 17 

shall be able to prepare emulsions of this type. It is also an impossibility to 
prepare filters that have negative transmissions in a given range. But in 
their paper, Hardy and Wurzburg indicate several methods whereby such 
a net result can be obtained. Of these methods, the masking process is the 
only one of interest to us. We shall have considerable to say concerning 
masking at a later time, but just now we will discuss their application of the 
technique. It is only the red-sensitivity curve that has a negative portion 
of sufficient moment to warrant reproduction. The complete red curve is 
therefore split up into two different parts, one of which will utilize a filter 
whose transmission will start at 535 m/z, rise gradually to a maximum at 610, 
then fall gradually to zero at 700 m/x. This we will call the "normal" red 
separation. 

In order to prepare the true red-sensitivity curve, the "normal" one must 
be corrected. This correction is made in the following manner. A filter is 
made which will have a transmission of zero at 400 mju, and which will rise 
to a maximum at 510 m/x, a maximum that will have a relative transmission 

VALUES FOR EQUATION 10 



Wave- 


_ 






Sr 


length 


X 


y 


z 


400 


O.OI 


0.00 


0.07 


O.OO 


410 


0.04 


0.00 


0.21 


O.OO 


420 


0.13 


0.00 


0.65 


O.OO 


430 


0.28 


0.01 


1.39 


O.OO 


440 


0.35 


0.02 


1.74 


O.OO 


45o 


0.34 


0.04 


1.77 


— 0.02 


460 


0.29 


0.06 


1.67 


— O.O3 


470 


0.20 


0.09 


1.29 


-0.05 


480 


O.IO 


0.14 


0.81 


— O.O7 


490 


0.03 


0.21 


0.47 


-O.08 


500 


0.00 


0.32 


0.27 


-O.08 


5io 


O.OI 


0.50 


0.16 


— O.O9 


520 


0.06 


0.71 


0.08 


-O.08 


53o , 


0.17 


0.86 


0.04 


-O.O3 


S4o 


0.29 


o.95 


0.02 


+ O.O4 


55o 


0.43 


1. 00 


O.OI 


O.I2 


560 


o-59 


1. 00 


0.00 


O.23 


570 


0.76 


0-95 


0.00 


0.34 


580 


0.92 


0.87 


0.00 


O.46 


59o 


1.03 


0.76 


0.00 


0.55 


600 


1.06 


0.63 


0.00 


O.S8 


610 


1. 00 


0.50 


0.00 


O.62 


620 


0.85 


0.38 


0.00 


O.49 


630 


0.64 


0.27 


0.00 


0.37 


640 


0.4s 


0.18 


0.00 


0.22 


650 


0.23 


O.II 


0.00 


O.16 


660 


0.16 


0.06 


0.00 


O.O9 


670 


0.09 


0.03 


0.00 


0.05 


680 


0.05 


0.02 


0.00 


O.O3 


690 


0.02 


O.OI 


0.00 


O.OI 


700 


O.OI 


0.00 


0.00 


0.00 



s b 


s e 


0.03 


O.OI 


O.II 


0.02 


o-35 


0.05 


0.74 


0.13 


0.91 


0.15 


0.94 


0.18 


0.92 


0.19 


0.68 


0.18 


0.43 


0.18 


0.25 


0.18 


0.14 


0.23 


0.09 


0.32 


0.03 


0.43 


0.00 


0.49 


O.OI 


0.52 


O.OI 


0.52 


O.OI 


0.49 


O.OI 


0.42 


O.OI 


0.24 


O.OI 


0.24 


0.00 


0.25 


0.00 


0.09 


0.00 


0.05 


0.00 


0.00 


0.00 


O.OI 


0.00 


O.OI 


0.00 


0.00 


0.00 


0.00 


0.00 


0.00 


0.00 


0.00 


0.00 


0.00 



i8 
/.o 

6 

•4- 
2 



4 ^ _ y Tl y , i g-jg 



HISTORY OF COLOR PHOTOGRAPHY 




580 6b© 420 646 660 6SO 700 



FIG. 4 




4O0 420 440 460 480 500 520 SAO 5*0 tfBO 60O 620 640 660 680 700 

FIG. 5 




400 420 440 460 480 S00 520 SAO SiO SSO 60O 620 646 460 480 700 

FIG. 6 

of 0.09 as compared to a value of 0.62 for the maximum transmission of the 
"normal" filter at 610. Beyond 510 it will fall, reaching zero at 535. This 
will then be the exact negative of that portion of the curve that has minus 
values for the transmission, the range from 400 to 535. Instead of making 
three negatives, four will be made: one each through niters whose transmis- 
sions correspond to S , and S&; a third through a filter whose transmission 
corresponds to the positive section of S r , which we termed the normal red 
filter; and a fourth through the filter which is the exact negative of the other 



SUBJECTIVE COLOR REPRODUCTION 19 

portion of S rj whose specifications we have just outlined. From this last 
exposure, a positive is made, and this positive will be superimposed in exact 
registry upon the normal red-separation negative. A little thought will soon 
show that the combination of the negative taken through the "normal" red 
filter combined with a positive of the negative taken through the correcting 
filter, will yield a new masked negative that will approach the desired S r 
sensitivity curve. 

In the last chapter it was pointed out that as far as the eye is concerned, 
it cannot tell whether a given light consists of a blend of hues or is mono- 
chromatic. Thus, in a colorimetric apparatus there will be no distinction 
between the light that is transmitted by the A filter, and monochromatic light 
of wavelength 610 m/x. In the same manner, it should be possible to match 
the light transmitted by a set of niters whose transmissions correspond to 
S r , S 0J and 5&. To determine the dominant wavelength and purity of each 
of these, we undergo a calculation identical to that made in Chapter 1 when 
the constants were determined for the various Wratten filters. The trans- 
mission of the filter is multiplied by the relative intensity of each wavelength 
in the light source, then by the tristimulus value for that wavelength. The 
results are summed over the entire spectrum to give the X, F, and Z values. 
From these the x, and y co-ordinates are determined, and the locus of the 
point corresponding to these values is plotted on the chromacity diagram. 
The dominant wavelength and purity can be read off from the diagram. When 
this calculation is carried out, using light of daylight quality for our source 
of illumination, the red filter, S r , has a dominant wavelength of 594.5 m/x 
and a purity of 100 per cent. The green filter, S , has a dominant wavelength 
of 520.5 m/x and a purity of 82 per cent. The blue filter, Sb, has a dominant 
wavelength of 468 m/x and a purity of 96 per cent. We compare three sets of 
filters, the standard A, B, and C5; the sharp-cutting 29, 49, and 61; the 
extreme set 50, 70, and 74; and the S r , S g , and S&. 





Red 




Green 






Blue 


A 


610 m/x 


B 


546 m/* 


85% 


C5 


46901/* 95% 


29 


631 


61 


543 


90 


49 


466 99 


70 


628 


74 


535 


97 


5o 


566c 92 


S r 


594-5 


s g 


520.5 


82 


s b 


468 96 



It is an extremely interesting phenomenon that of all the filter sets illustrated, 
the standard set approaches the ideal subjective curves most closely, but not 
sufficiently to be considered duplicates of them. It must be recalled, however, 
that the agreement cannot be too exact, as the ideal curves were developed 
upon the assumption that the product of the light intensity with the spectral 
sensitivity of the emulsion is a constant throughout the entire range from 
400 to 700 m/x. This is only a very rough approximation. 

Consider a colorimeter that contains filters whose transmissions correspond 
to those of S r , S , and Sb- Such a colorimeter, of course, is an ideal that can- 



20 HISTORY OF COLOR PHOTOGRAPHY 

not be achieved, since it is not possible to make filters that have negative 
transmissions. Now compare this colorimeter with one in which the filters 
transmit monochromatic colors as follows: The red filter will transmit light 
of wavelength 594.5 m/x, the green filter will transmit light of wavelength 
520.5 mjit and a purity of 82 per cent, and the blue will transmit light of wave- 
length 468 mju and a purity of 96 per cent. As far as the eye is concerned, 
it would be impossible to distinguish between the two sets. 

In this chapter we have adopted the point of view that color photography 
is merely applied colorimetry. Color analysis should then be an attempt to 
duplicate the mechanism by which the eye analyzes color; that is, in terms 
of filters such that the final result would be a spectral sensitivity that corre- 
sponds to the shape of the curves S r , S , and Sb. It was indicated that as far 
as the eye is concerned these could not be differentiated from monochromatic 
filters whose transmissions are 594.5, 520.5, and 468 m/x respectively. Color 
photography in accordance with the subjective principles should be closely 
approached, therefore, if the analysis filters are made to correspond to these 
transmissions. This is an entirely different matter from the preparation of 
filters that correspond to S r , S g , and S&. It should not be a difficult matter 
to prepare filters which would correspond to the above specifications. In 
fact the C5 filter is a very close approach to the proper blue. It must be borne 
in mind that it is not necessary to make true monochromats, for the eye 
cannot tell whether a given light is really monochromatic or whether it is a 
blend of many other colors. The red filters are very good examples of this, 
for as far as the eye is concerned these are true monochromats. From this 
point of view, it is also immaterial whether the range of transmissions of the 
three filters are mutually exclusive, or whether they overlap considerably at 
the edges, the only important thing being that the blend of the total trans- 
missions must correspond to the specifications outlined above. From the 
point of view of photography this means that broad bands could be used, a 
very useful concession, since this would allow greater speeds. 

In additive processes, the same filters are used for taking and viewing. 
Thus in the line-screen and starch-grain processes the primary colors form a 
component layer of the sensitive material, and this is processed by reversal. 
Hence the same set of primaries is used in the final image. The above specula- 
tions are then at once applicable. The starch grains, or the different lines 
that make up the filter layer, must be dyed to correspond to the specifications 
above. It will be shown in a later chapter that by means of accurate proc- 
essing it is possible to make the silver deposit correspond identically to the 
"density" of light that is reflected from the corresponding point in the original 
image. 

But additive processes have a very limited application at the present time, 
since they are restricted to projection. The application of the above ideas 
to the subtractive processes is not so simple. In these procedures the positive 
that is made from the negative that is exposed behind the red filter, must be 



SUBJECTIVE COLOR REPRODUCTION 21 

• 

converted into a color which will absorb the red primary and transmit com- 
pletely the green and the blue. This in itself is not a very difficult matter 
if the densities are quite heavy, but when light densities are to be reproduced, 
then it will be found that the spectral absorption of the color will not be nearly 
as sharp-cutting. This means that unless truly monochromatic colors are 
used in the synthesis scheme, the characteristics of the color will change as 
the density changes. It may very well be that this broadening out of the 
transmission band as the intensities become lower, is of very little moment; 
perhaps the increase in the range on one side is exactly balanced by the 
increase on the other side of the band, but this is a matter that has never been 
tested. 

The chromacity diagram can be utilized to determine the exact shade into 
which the positives made from the three separation negatives, must be con- 
verted. The positive is the exact complement of the negative, therefore the 
color into which it must be converted must be the exact complement of the 
negative filter color. In the previous chapter it was shown that the comple- 
ment of any color can be obtained by the mere expedient of determining the 
point of intersection of the line from the color under consideration, to the 
point which designates white, with the curve of the spectral colors. Any 
straight line through the point (Fig. 2, p. 11), representing white, must 
intersect this curve in two places. These two points represent complementary 
colors. Thus the A filter is represented in the chromacity diagram by the 
point which represents monochromatic light of wavelength 610 mju. To 
determine its complement, this point is connected with the point 0, and the 
line extended until it intersects the spectral curve again. This will happen 
at the point 490 m/z, a blue-green. In the same manner we can determine the 
colors complementary to the set of ideal filters S r , S g , and St>. 



Filter 



Dominant Complementary 

Wavelength Wavelength 

S r 594.5 mix 487.5 mju, a blue-green or cyan 

S g 5 2 °-5 520.5c, a purple or magenta 

Sb 468 57 2 -5> a yellow 

When we discussed the taking filters we pointed out that the detailed trans- 
mission of the filters was of no account, provided the complete blend had the 
same co-ordinates in the chromacity diagram that the filters had. This is 
because the eye cannot distinguish a blend of colors from a pure one. This 
condition no longer holds for the secondary or complementary colors. This 
is made clearer by an examination as to their exact function. Let us consider 
the positive made from the red-filter negative. This filter transmits the 
equivalent of monochromatic light of wavelength 594.5 mju. The positive 
must therefore deposit densities which will absorb this wavelength to an 
extent that is proportional to the silver deposit at each point. But under no 



22 HISTORY OF COLOR PHOTOGRAPHY 

• 

circumstances must it deposit densities which will absorb any of the other 
two primary colors. Therefore the colored positive must transmit freely the 
wavelengths 520.5 and 468 mju. It is no longer sufficient for the secondary 
color to have the right hue, but it must transmit the other two primaries 
completely. The two primaries must be so blended that their combination 
will be matched by monochromatic light of wavelength 487.5. The positive 
made from the green negative must be colored so that it will transmit freely 
the equivalents of wavelengths 594.5 and 468. This will be a purple, matched 
by a color whose designation is 520.5c. 

It is very unfortunate that color photography in accordance with the sub- 
jective principles has never yet been really tried out. And this despite the 
fact that the very first attempts to solve the problem of color reproduction had 
their foundations laid directly upon the Young trichrome theory, a hypothesis 
which is completely subjective. Ives reformulated these ideas in his early 
patents, but still the practice continued to rely upon the purely objective 
mechanism. This may be due to the fact that it is only quite recently that 
the subject of colorimetry has been advanced sufficiently to be able to give 
quantitative values to the various constants. Now that the subject has been 
so thoroughly aired, it is to be hoped that experimental verification will soon 
follow. To those who are desirous of following this idea further, it is recom- 
mended that the following papers be studied in the originals. 

J. C. Maxwell, Trans. Roy. Soc. Edin., Vol. 21 (1855) p. 275. 

F. E. Ives, /. Frank. Inst., Vol. 125 (1888), p. 345; Vol. 127 (1889), p. 54. 

E. Schrodinger, Lehrbuch der Physik (Muller-Pouillets), II, Optics, Part 1, p. 488. 

C. Schaefer & K. Ackermann, Zeit. tech. Physik, Vol. 8 (1927), p. 55. 

G. B. Harrison & R. G. Horner, Phot. J., Vol. 77 (1937), p. 706. 
H. Frieser & R. Reuther, Zeit. tech. Physik, Vol. 19 (1938), p. 77. 
H. D. Murray & D. A. Spencer, Phot. /., Vol. 78 (1938), p. 474. 

D. L. MacAdam, /. Opt. Soc. Amer., Vol. 28 (1938), p. 399. 



CHAPTER 3 
OBJECTIVE COLOR REPRODUCTION 



I 



N the last chapter we discussed color reproduction from the subjective 
point of view. In that procedure the attempt is made to duplicate the processes 
by which color sensations are registered by the brain. Now we- will discuss 
objective color reproduction, where the attempt is made to duplicate the 
processes which cause objects to appear colored. This is purely a physical 
phenomenon, whereas the former is purely psychological. 

The reason why any object appears colored, rather than a neutral gray, is 
that the object has a selective absorption or reflection of the light that is inci- 
dent upon it. An object will appear to be colored yellow if it will absorb the 
blue and transmit the red and green rays. It is possible to draw a curve which 
gives the percentage transmission of each wavelength that is reflected from a 
colored object. This curve is called the spectrum of the color. In this case 
we no longer consider color to be a sensation sensed only by some mysterious 
forces within the human brain, but we must consider it a property of matter, 
the property of selective absorption of incident light. In this definition we 
can identify the color absolutely by its spectrum. Two colors whose spectra 
are identical, will be identical under all conditions. If it be required to repro- 
duce a given color, it is merely necessary to reproducer color with an identical 
spectrum. This is the true aim of objective color reproduction. 

There have been proposed many methods whereby the spectrum of a given 
color or set of colors can be duplicated automatically and precisely. The 
Lippmann process of interference photography is one elegant solution that 
received considerable attention at the beginning of the twentieth century. 
This process was first suggested or hinted at by J. Nicephore Niepce (1829), 
but it was not until G. Lippmann used the grainless albumen emulsion de- 
scribed in Eder's Handbuch (Vol. 2, p. 533), that any success was achieved. A 
grainless sensitive emulsion is coated upon a glass plate, and this is brought 
in contact with a highly polished mirror surface. When light waves pass 
through the glass plate and the emulsion and are reflected back by the mirror, 
the reflected light, will, in accordance with the principles of physical optics, 
be out of phase with the incident light by an amount equal to half a wave- 
length. The reflected and the incident light, traveling in opposite directions 
and being out of phase by one-half of a wavelength, will form stationary waves. 
These will be characterized by the fact that at the nodal points, situated half 
a wavelength apart, the two waves will just neutralize each other, and at these 

23 



24 HISTORY OF COLOR PHOTOGRAPHY 

points the light intensity will be zero. At the crests, also situated a distance 
of half a wavelength apart, the light intensities will add to each other, and, 
at these points, the intensity will be doubled. If this takes place within the 
confines of an emulsion, then at the positions of the crests, there will be 
formed latent images, and after development, silver laminar deposits, capable 
of acting as mirrors. At the other points, there will be no photochemical ' 
action. Each wavelength will deposit its own system of mirrors situated half 
a wavelength apart. When illuminated with light, each of these systems will 
reconstruct the original wavelengths, hence the original colors. But the il- 
lumination of these plates must be in accordance with the taking scheme; 
that is, with a mirror in contact with the print so that standing waves will be 
reformed. For this reason, the plates can be viewed only at a certain critical 
angle. The process is very slow, because the use of grainless emulsions is 
necessary. These are about 10,000 to 1,000,000 times as slow as ordinary 
negative materials. For these reasons the Lippmann process has remained 
the plaything of the scientists. The one very great contribution that work in 
this field has achieved, is the preparation of grainless emulsions that may be 
useful in fields more practical than interference photography. Their prepara- 
tion is not essentially difficult. H. E. Ives (Brit. J. Phot., Vol. 55 (1908), 
pp. 942, 965, 979; Vol. 56 (1909), Color Sup., p. 7), discloses the following 
procedure: 

Solution A Solution B 

25 parts Water (distilled) 50 parts Water 

1 part Gelatin i part Potassium bromide 

2 parts Gelatin 

Solution C 
5 parts Water 
0.3 part Silver nitrate 

Solutions A and B are warmed until the gelatin is melted, then they are cooled 
to 40 C, at which temperature solution C is added to A, and the mixture 
then added to B. After the addition of the sensitizing agents (pinacyanol, 
pinaverdol, cyanine, erythrosine, etc.), the emulsion is filtered, coated upon 
glass plates, washed for fifteen minutes in running water, then dried. R. E. 
Liesegang (Kol. Zeit.,Vo\. 17 (1916), p. 36), disclosed a very convenient method 
for the preparation of such plates. Glass plates are coated with 6 per cent 
gelatin solution containing 3 per cent KBr. As soon as the gelatin has set, 
the plates are bathed for approximately five minutes in a 5 per cent solution 
of silver nitrate, after which they are washed with distilled water. The optical 
sensitization of these plates is not very easy. It seems that only erythrosine 
and cyanine could be used. Otherwise this method would be an extremely 
simple one. 

Another ideal solution to the problem of objective color reproduction is 
the "micro-dispersion" method. This was first disclosed by F. M. Lanchester 



OBJECTIVE COLOR REPRODUCTION 25 

(Eng. P. 16548/95). In this scheme (Fig. 7), a lens L\ projects an image upon 
a grating G } which consists of ruled parallel opaque lines, approximately 300 
to the inch, with the width of the opaque lines about twice the width of the 
clear spaces between them. This image is then projected by a second lens, 
L2, on to a screen. Between the second lens and the screen there is placed a 
prism whose angle of approximately 2 degrees, is sufficient to disperse the 
light traversing it into minute spectra that will just fill in the space between 
the images of the clear spaces in the grating. The axis of the prism is parallel 
to the lines in the grating. 



Prison, 






vcoA 



?ct 



FIG. 7 

By this scheme, the light that is reflected from the object is brought to a 
focus in the plane of the grating. This acts as a new original, composed of 
minute lines. The second lens will now image each one of the lines at a dif- 
ferent point on the screen. When the image of one of the lines enters the 
prism, it is dispersed into a spectrum which is a true and absolute specification 
of the color of the light reflected from the corresponding point of the original. 
If the grating contains 300 lines to the inch, and if the second lens system 
magnifies the image three times, then the screen will contain 100 micro spectra 
to the inch, and the image on the screen will appear slightly coarser than the 
images in an ordinary magazine illustration, which breaks up the image into 
133 units per inch. Now let us replace the screen by a panchromatic emulsion, # 
and give this a balanced exposure. The negative, after development, will 
yield silver deposits whose densities will follow the spectral distribution of the 
light. Let us suppose that we are photographing a pure red light, consisting 
of all the wavelengths greater than 600 m/i. p The negative will in that case 
have silver deposits only in that portion of each minute unit of 1/100 inch, 
which will correspond to the position of the red portion of the spectral band, 
say the upper third. The complete negative image will then consist of a 
series of lines, each one-third of 1/100 of an inch long, and running 100 to the 
inch. Let us now place this negative back into the original position, and this 
time project white light upon the grating, instead of colored light. The 
second lens will project this upon the screen, but upon the passage through 
the prism, the white light will be dispersed into a prism. In the position 



26 HISTORY OF COLOR PHOTOGRAPHY 

formerly occupied by the screen, we have placed our negative with its dot-like 
image. When this is synchronized with the grating, the image will just blot 
out that section of the minute spectra which will correspond to the wave- 
lengths that are greater than 600 mju, hence the light that will emerge from 
the other side of this negative will be the exact complement of red. 

Instead of processing the original to form a negative, it can be converted 
into a positive, taking great care that sizes are maintained in a strict one-to- 
one ratio. If such a positive be inserted in the position in exact registry with 
the image of the grating, then the color that will emerge will be red. 

In an independent manner, and from a slightly different point of view, 
G. Lippmann {Brit. J. Phot., Vol. 53 (1906), p. 644), whose contributions to 
color we have already discussed in this chapter, was led to formulate ap- 
proximately the same scheme. Other workers who have disclosed similar 
or modified ideas, were P. E. B. Jourdain {Brit. J. Phot., Vol. 46 (1899), 
pp. 232, 262), A. F. Cheron (Fr. P. 364,526), and M. Raymond {Photo-Revue, 
Vol. 19 (1907), p. 51), and J. and E. Rheinberg. These last used a line screen 
with 500 lines to the inch {Phot. J., Vol. 52 (191 2), p. 162). The clear spaces 
were three times as wide as the opaque. A direct vision prism giving a normal 
spectrum was used. Transparencies made from the negative had to be viewed 
in the taking apparatus. The Raymond disclosure appears to be of interest, 
as it indicates the use of a compensating prism in the viewing apparatus. A 
lens, I* (Fig. 8), projects an image of the object upon the grating G, which 
is similar to the one used by Lanchester. This image, consisting of lines a 
fraction of a millimeter in width, was then projected by another lens L% at 



I p/ajte 



Prt$*K 




" for 1/L Q.H/LtU? 



FIG. 8 



a one-to-one ratio upon the sensitive material. Interposed between the 
lens L2 and the emulsion was an appropriate prism to yield micro spectra 
that will just cover the spaces between the lines. Positives were made from 
the negative and positioned in the place formerly occupied by the negative. 
To view this, a compensating prism was placed between the eye and the 
positive. 

The idea seems to have been dormant for some time. It was resurrected 
again by S. Prisner in 1931 (Eng. P. 341,078). In this scheme (Fig. 9), 
the light from a lens L\ goes to a special lens L2, which parallelizes it into a 



OBJECTIVE COLOR REPRODUCTION 27 

narrow beam, and causes it to fall upon a prism. The lens L2, must obviously 
be of sufficient aperture to cover the entire image. The sensitive plate is 
placed with its emulsion side in contact or adjacent to the emergent side of 
the prism. For projection, the positive is placed between two prisms situated 
complementary to each other, and the entire assembly placed in the same 
relative position formerly occupied by the prism and the negative emulsion. 
F. Preinerstorfer (U.S.P. 2,088,399) disclosed an idea which appears to be a 
combination of the original Lanchester disclosure with that of C. Urban 
(Eng. P. 8723/07; Fr. P. 376,616). This last consists in placing the grating 
behind the second lens. 





FIG. 9 

The micro-dispersion methods of direct color reproduction are mere labora- 
tory curiosities, principally because they are extremely slow, and because they 
call for very special apparatus both for taking and for viewing. In order 
to obtain satisfactory sharpness, the image must be broken up into at least 
five hundred lines to the inch. Thus each micro-spectrum must be less than 
1/500 inch, and, in this extremely small space, there must be condensed the 
image of a spectral range of 300 mju. A pure yellow would correspond to a 
silver deposit that covers only one-third of this area, so that the emulsion 
would have to be capable of resolving 1500 lines to the inch. I do not believe 
that many existing fast negative materials could give that resolution, and it 
must be remembered that this is but the minimum. Of course with grainless 
emulsions of the Lippmann type, this does yield an absolute solution to the 
problem. 

It appears to me that the micro-dispersion methods may be utilized to 
prepare filters that have specified transmissions. In this case grainless emul- 
sions can be used. ' Let us suppose that we desire to prepare a filter that will 
transmit 100 per cent of the remaining light. Then we set up a micro-dispersion 
system and photograph a white area, say a white opal glass. We illuminate 
this glass from the opposite side in the following manner. A source of light 
is dispersed into a broad spectrum by means of a wide-angled prism, and upon 
the emergent side of this prism is placed a template which will cut out all the 
undesired wavelengths. The template is made the "sandwich" filling be- 
tween two prisms that are placed complementary to each other. The emer- 
gent light is projected upon the opal glass, and it is photographed to give a 
master from which duplicates can be made. These duplicates are then made 
the filling of a sandwich made of two complementary prisms identical to the 



28 HISTORY OF COLOR PHOTOGRAPHY 

one used in the photography. On the incident side of one of the prisms, there 
is placed a grating that is identical with the projection of the grating used 
in the original photography upon the prism when this is in the focal plane 
of the second lens. The completed sandwich will then act as a filter when 
illuminated by collimated light. Under ordinary circumstances, such filters 
would have very little application, and their preparation would merely serve 
as a convenient discipline. But under certain conditions, as for example in 
investigations of a purely theoretical nature as to the effect of sharp-cutting 
or of widely overlapping filters upon reproduction processes, it may be essen- 
tial to be able to prepare such ideal theoretical filters. It may be possible to 
prepare in this manner filters whose transmissions would correspond to S r , 
S g , and Sb. The masking process can be applied to the preparation of S r in 
this case just as easily as in the case of ordinary photography, although 
much greater care must be exercised to maintain accuracy. 

The procedures outlined in this chapter up to this point are of only scientific 
interest, since their use involves conditions that are beyond the scope of or- 
dinary practice. The successful application of objective principles that un- 
derlie methods of color reproduction, deals with attempts to approximate the 
spectral curves of the original colors, rather than with attempts to duplicate 
these curves exactly. The technique adopted to accomplish this is identical 
with the devices adopted by the mathematicians when they desire to deter- 
mine the area under a complicated curve. The analogy is so close that it may 
be well to demonstrate the procedure in greater detail. Let us consider the 
spectral curve of the No. 30 Wratten filter, in the range from 400 to 700 m/z. 
The transmissions are listed in the table below: 



400 


63.0% 


500 


0.0% 


600 


81.2% 


410 


61.4 


5io 


0.0 


610 


84.5 


420 


57.6 


520 


0.0 


620 


86.0 


430 


52.8 


53o 


0.0 


630 


86.7 


440 


45-8 


54o 


0.0 


640 


87.4 


4SO 


38.1 


55o 


0.0 


650 


87.7 


460 


29.5 


560 


4.0 


660 


88.0 


470 


18.2 


570 


31.6 


670 


88.0 


480 


8.0 


580 


63.1 


680 


88.0 


490 


1.3 


59o 


75.5 


690 


88.0 



The curve is shown in Figs. 10, n, 12 and 13. The energy associated with the 
light transmitted by this filter is proportional to the area under the curve, 
and for that reason it is of interest to determine this value. Incidentally 
the photographic effect of the light is directly proportional to the light energy. 
In order to determine this area, the mathematician will divide the curve into 
a series of rectangles whose upper side will be a straight line that is some 
average of the line that makes up the curve. The area under the curve will 
then be the sum of all the areas of the rectangles into which the curve has 
been subdivided. The area of each rectangle will be equal to the product 



OBJECTIVE COLOR REPRODUCTION 



29 





400 450 500 550 600 650 700 

SPECTRUM OF J/o.30 WRATTE/V FILTER. 
FIG. 10 




400 



450 500 550 600 

30-COLOR ANALYSIS 

FIG. 11 



700 



xx x x 




4oo 



450 500 550 600 
2- COLOR ANALYSIS 



650 700 



FIG. 12 



of the width by the height, and if the first division is made into thirty equal 
parts, each io m/z in width, then the area will be ten multiplied by the value 
of the average transmission of the filter in that io mju range. The rectangle 
bounded by the values 400 and 410 m/x will have an average transmission of 
i(63 + 61.4) or 62.2 per cent, so its area will be 622 units. The areas in the 
other rectangles will be found in the same manner, and the sum will be a 
measure of the total area. Obviously the smaller the width of the rectangles, 



3° 



HISTORY OF COLOR PHOTOGRAPHY 



the closer will the approach be to the true value. The absolute value will 
therefore be obtained only when the curve is divided into an infinity of rec- 
tangles each innnitesimally wide. When this stage is reached, the mathema- 
tician says that he "integrated" the expression ydx, which represents the 
area under the curve if y represents the relative transmission of the filter, 
and x represents wavelength corresponding to the transmission. 



f ^* - *** x y yyyv 




400 450 500 550 600 650 
3- COLOR ANALYSIS 



70O 



FIG. 13 



It was stated above that the area under the curve is directly proportional 
to the energy associated with the light transmitted by the filter. This energy 
is also a measure of the photographic effect that the light transmitted by the 
filter will have upon an emulsion. When the photographer exposes a sensitive 
plate behind a filter, he is really making a determination of the light energy 
that the filter transmits. Hence a measure of the silver density produced 
by such an exposure is a direct measure of the area under the curve, and can 
be substituted for it. In order, then, to determine photographically the area 
under the curve, it becomes necessary to prepare a set of filters which will 
transmit only selected portions of the spectrum. If the analysis is to be made 
in terms of thirty elementary rectangles, then each filter must transmit com- 
pletely the light in narrow bands of 10 m/i width, and absorb completely 
the remaining light. The light whose curve is to be determined is then 
brought into the lens of the camera, over which is placed these filters one 
after the other. Thirty exposures are required, timed to make up for the 
different relative sensitivities of the emulsion to the light passed by a particu- 
lar filter. The plates are then all developed to the same contrast, and the 
densities measured. These values will be directly proportional to the relative 
transmissions in the filter, so that the thirty points will reconstitute the curve 
of the original color. Obviously to duplicate the curve exactly, there will 
have to be an infinity of such exposures through an infinity of filters, each 
of which transmits pure monochromatic light. 

When the problem of objective color reproduction is restated in these terms, 
the question can be immediately asked into how few bands can the curve be 



OBJECTIVE COLOR REPRODUCTION 31 

divided so that the synthesized curve approaches the true one within the 
limits of accuracy attainable in reproduction processes. Here eye tolerance 
plays a very large role. This is the only application of psychological processes 
that is used in the objective schemes. This is made use of in the following 
manner. From each of the negatives prepared in the manner disclosed above, 
there is made a positive, so processed that the transmission through the posi- 
tive will be identical to the transmission in the original filter, but this trans- 
mission is color blind. This can be done very easily, as will be made evident 
in Chapter 7. Each of these positives is then illuminated with the light that 
is transmitted by the filter behind which the corresponding negative was 
exposed, and the transmitted light is then projected in registry upon a white 
screen, or is led into the same half of a colorimetric apparatus. The other 
half is illuminated by the original light, and the two compared. The problem 
can now be stated as into how few bands must the color be analyzed in order 
that the eye will not be able to distinguish the duplicate from the original. 

The smallest number of rectangles into which any curve can be divided is 
two, leaving out the trivial case of a single rectangle. To carry out this 
experiment, two filters must be prepared, the first of which transmits the 
light up to 550 m/x; the second transmits the light from 550 m/z on. The 
transmissions of the two filters must be mutually exclusive. When these 
requirements are met, and the colors compared in the colorimeter, it will be 
found that a surprisingly good approach to the truth will be obtained. There 
will be no inversion of colors; that is, there will be no cases where a red color 
would become duplicated as a green or blue, or vice versa. But no single color 
will be accurately reproduced. 

The next simplest approach will be that of a three-band analysis. In this 
case the filter transmissions will be 400-500, 500-600, and 600-700 mju. 
The hue of the first band will be a blue-violet, that of the second a pure green, 
and that of the third a red that verges upon the orange. When colors so 
duplicated are compared with the originals in a colorimeter, the approach 
will be found to be just within the limits of differentiation. It will be only 
when the original is placed directly beside the duplicate that a difference 
will be found between them. This very close correspondence, coupled with 
the hypothetical speculation of the Young-Helmholtz theories, gave rise to 
the belief that the one was an experimental verification of the other. But 
this is entirely an illusion. In the first place, the duplication was not made 
in accordance with the requirements of the subjective hypothesis, that is, 
in terms of filters whose transmissions correspond to S n S , and Sb. The two 
sets of filters differ very much, in that the ones suitable for the subjective 
processes overlap considerably, while the others are mutually exclusive. It 
may very well be that the dominant hues of the two sets of filters are not very 
far apart, as is seen from the table at the end of Chapter 2. It is this close 
approximation of the dominant hues of the two sets of filters that has given 
rise to so much loose talk, for when the different experimenters talked glibly 



32 HISTORY OF COLOR PHOTOGRAPHY 

of the three-color theory of reproduction as being based upon the Young- 
Helmholtz hypotheses, they utilized the mutually-exclusive niters for their 
experiments. Hence their experimental work was not a test of their specula- 
tions — a fact realized very clearly by Hardy and Wurzburg in the paper 
detailed in the last chapter. 

The very close approach which a three-band analysis makes to the original 
color makes speculation concerning a four-color analysis extremely interesting. 
The filter transmissions now must be 400-475 m/z, 475-550 m/i, 550-625 mju, 
and 625-700 mju. These colors are a violet, a bluish green, an orange-yellow, 
and a deep red, respectively. As far as is known this scheme has not yet 
been tested, but from the results of the three-color analysis, it may be inferred 
that the duplication of the colors by this system would be well beyond the 
ability of the eye to differentiate. 

So far we have discussed the analysis colors in color reproduction. The 
synthesis colors in objective systems must fulfill certain definite requirements 
as to transmissions and absorptions. In the additive systems, the synthesis 
and analysis colors are identical, but in the subtractive systems they are 
complementary to each other. This was discussed slightly in the last chapter, 
but the same things are true here. The positive print made from the negative 
that is exposed behind the red filter must be converted into a color that will 
absorb the red primary in proportion to the density of the image, and it must 
transmit completely the other two primaries. Such a color will be a cyan, or 
blue-green. Similarly the blue-filter positive must be converted into a yellow 
color which will transmit the red and the green, and absorb the blue. The 
green-filter positive must be converted into a magenta hue, which will trans- 
mit the red and the blue, but which will absorb the green. Despite the high 
state of perfection of the dye industry, no dyes are known at the time of this 
writing which completely satisfy these requirements. In fact, it is only the 
yellow colors that even approach the necessary transmissions to an extent 
of approximately 85 to 90 per cent. The other two colors are much worse. 
The blue-greens especially are poor in that they satisfy theory only to an 
extent of less than 50 per cent. The magentas are slightly better. Most of 
the very poor results in subtractive processes are due to this lack of satis- 
factory secondary colors. Some corrective measures have been disclosed in 
a procedure termed masking, but we will leave a discussion of this to Chap- 
ter 19. The ideal transmissions for the secondary colors in a three-color 
analysis scheme are listed in Fig. 14. 

When the analysis of the color is made in terms of four bands, then the 
secondary colors must transmit three of these bands and absorb but one. 
The preparation of the secondary colors for a three-color analysis is far from 
perfect. The four-color secondaries would present a problem that is of a 
higher order of difficulty than the other. Since no four-color color analysis 
has ever been attempted, there has not yet arisen the need for such secondary 
colors. 



OBJECTIVE COLOR REPRODUCTION 



33 



8 



40 



* 



2(h 



400 



5*00 600 

/deal Yellow 



700 



wye 




1 soo t u 60 ° 

/deal Magenta, 



700>KfL 




IcUzq] Cyan, 



600 



700 >nus 



FIG. 14 



The results outlined above can be generalized. Let us suppose we make 
our analysis in terms of n bands, which divide the spectrum into n equal 
divisions each 300/re m/i in width, and each of them mutually exclusive. 
As the number n increases, the need for exclusiveness.is greater. The corre- 
sponding secondary color will in each case absorb one band that is 300/7? m/i 
in width, and will transmit completely the remaining n — 1 bands. 



CHAPTER 4 

COLOR CAMERAS 

J/ROM the discussion in the previous chapters it is evident that the first 
requisite for color reproduction is the preparation of color separations. When 
the subject matter is immobile, it suffices to photograph the object three 
times, each time through a filter which transmits but one of the primaries. 
In this manner three negatives are obtained such that each one of them repre- 
sents the intensities of each primary at every point of the object. But when 
the subject matter is such that it is not possible to effect this multiple pho- 
tography, then the three exposures must be made simultaneously. This can 
be done by means of a one-shot camera. 

Many variations of such cameras are possible, but in all of them there is 
one common factor. The light that is reflected from the object into the 
camera is divided into three sub-beams, and each one of the sub-beams is 
directed to a negative material that registers but one of the three primaries. 
It must be understood that the term primary is being used in a general sense, 
and not with the connotation current in subjective processes. By the term 
primary is meant the color or beam of light that is transmitted by one-third 
of the spectrum. The blue primary will therefore denote the ensemble of 
colors that have their wavelengths in the range between 400 and 500 mjj,, 
the green primary will lie in the range between 500 and 600 m/z, and the 
red primary will lie in the region bounded by 600 and 700 m/x. Cameras 
differ from each other only in the design of this light-splitting mechanism. 

The ideal camera will consist of a lens in combination with a light-splitting 
device which has the following properties. Let us suppose that a beam of 
white light enters the camera. It proceeds to the light-splitting mechanism 
which divides the beam into three sub-beams, and each of these is deflected 
to a different negative material. The ideal camera will so divide the main 
beam that all the red rays will constitute one of the sub-beams; all the green 
rays will constitute the second sub-beam; and all the blue rays will form the 
third sub-beam. Such a light-splitting device will lose no light due to the 
use of light filters, substances which lose at least 70 per cent of the incident 
light in three-color processes. Such an ideal camera is not an impossibility. 
It was disclosed by some scientists that by deposition of very thin layers of 
substances that have very high indices of refraction for certain colors with 
layers that have very low indices of refraction, it becomes possible to make 
reflectors which will transmit certain primary colors with great efficiency, 

34 



COLOR CAMERAS 35 

approximately 90 to 95 per cent, and which will reflect other primaries with 
equally great efficiencies. Thus it will become possible to make reflectors 
which will transmit 90 to 95 per cent of the red and green primaries, and 
which will reflect 90 to .95 per cent of the blue. Our ideal camera will contain 
two such reflectors, one which will reflect the blue rays, the other which will 
reflect the green rays. When a beam of white light enters the camera, the 
first reflector will transmit 90 per cent of the red and green primaries to the 
second. It will reflect 90 per cent of the blue light. There is thus a loss of 
only 10 per cent in light efficiency by the action of the first reflector. When 
the transmitted light reaches the second reflector, it will reflect 90 per cent 
of the green rays so that 80 per cent of the original green light will reach 
the green-sensitive negative material. This will also transmit 90 per cent 
of the remaining red rays so that 80 per cent of the original red rays will be 
deflected to the red-sensitive material. Since 90 per cent of the original blue 
rays will be effective, such a camera will have a light efficiency of over 80 per 
cent. It is permissible therefore to rate color cameras in efficiency, relative 
to this ideal. 

The various light-splitting devices used in color cameras were classified 
by Adrian Klein in a very effective manner. In this discussion we will follow 
his classification, with a slight change in the order of numbering. This last 
is done to facilitate the discussion, rather than to attempt to improve his 
nomenclature. There are five main classes, of which two are of academic 
interest only. We will treat these first, so we designate them as Types 1 and 2. 

1. Two or more lenses without beam division. 

2. Two or more lenses behind inclined glass plates. 

3. One lens in front of beam splitters. 

4. Two or more lenses behind beam splitters. 

5. A divergent lens in front of two or more lenses. 

Before we begin our discussion concerning the various types enumerated 
above, it may be well to discuss briefly the elementary principles regarding 
lenses. When a beam of light passes from one medium (air), into another, 
which has an index of refraction that is different from the first, the direction 
of the beam is changed. The extent of this change is determined by the curva- 
ture of the interface between the two media and by the difference in the indices 
of refraction. From this fundamental fact, it is an easy matter to demonstrate 
exactly why the rays of light coming from a very distant point are brought 
to a point focus by means of a lens both of whose sides are convex. Such 
lenses are called convergent or positive lenses, and the ordinary lenses used 
in cameras are of this type. It is possible to grind the surfaces of a lens so 
that the beam of light that enters the lens is not brought to a point focus, 
but is spread out into another beam of light that is divergent. Such lenses 
are termed divergent or negative lenses. Both of these types are useful in 
camera design. 



36 HISTORY OF COLOR PHOTOGRAPHY 

In photography, the lenses commonly used have a focal length that varies 
from approximately one inch to twenty or more inches. In motion-picture 
work, and in candid cameras, where motion-picture film is used, the focal 
length of the lens is usually about 50 mm, or roughly two inches. This means 
that a beam of light that originates at infinity (relatively) will be brought to 
a point focus two inches behind the rear nodal point of the lens system. In 
still photography, when 5 by 7 inch film negatives are used the lens has a 
focal length of approximately eight inches. This means that the beam of 
parallel rays will be brought to a point focus exactly eight inches away from 
the rear nodal point of the system. 

Another characteristic of lenses is the aperture. This is a function of the 
effective diameter of the lens. It is customary to designate this property by 
a number like 7:4.5, /: 6.3, etc. This is the ratio that the effective diameter 
of the lens bears to the focal length, so that 7:4.5 means that the focal length 
of the lens is 4.5 times the effective diameter. Consider two lenses identical 
as far as structure is concerned, but differing only in / number; that is, one 
has a rating of 7:4.5, while the other has a rating oiflg. This means that the 
effective diameter of the lens in one case is exactly twice that in the other. 
How much light will enter the two lenses? The amount of light that enters 
a camera lens is determined by the effective diameter of the lens. Suppose 
that the two lenses are of nine-inch focal length. The effective diameter in 
the case of the 7:4.5 lens will be two inches, while in the case of the 7^9 lens 
it will be only one inch. The amount of light that will enter a lens is directly 
proportional to the area of the lens that is exposed to the beam of light. This 
is proportional to the square of the diameter, so that in our case it becomes 
evident that the 7-4-5 lens will receive four times the light that will be re- 
ceived by the f: 9 lens. 

Let us now consider what happens when we photograph an object situated 
about 25 feet away from the optical center of the lens system, with a camera 
that is equipped with a 9-inch focal length lens operated at an aperture of 
7:4.5. From every point on the object there will originate a cone of light, 
which, when it reaches the camera lens, will have a cross section equal to a 
circle with a diameter of two inches. The axis of the cone will be 25 feet in 
length, so that its angularity will be very slight. To all practical purposes 
we can disregard the fact that the beam of light is a cone, and consider it to 
be a bundle of parallel rays emanating from infinity. This bundle of rays 
will enter the lens of the camera at a definite inclination to the optical axis 
of the lens. Since the bundle is to be considered as consisting of parallel rays, 
it will be brought to a focus at a point in the plane which is exactly nine inches 
behind the optical center of the lens, and perpendicular to the optical axis. 
This is the plane in which we place the negative material. The exact position 
to which the rays will be focused is determined by only one fact, the angle 
which its axis makes with the optical axis. Two points on the object, that 
are separated by a very small amount, will generate two beams of light that 



COLOR CAMERAS 37 

will enter the lens system at different angles, hence the two beams will be 
brought to a focus at different points in the focal plane of the lens. This 
is a very important characteristic of lenses, arid upon this property depends 
the entire system of lenticular processes. 

We can now make short shrift of the devices that are classified in types (1) 
and (2). Consider the beam that originates from a point on the object, and 
which enters two lenses situated side by side, with their optical axes separated 
by a small amount. Each lens will direct the beam entering it, to a negative 
material which is sensitive to but one primary. For the purposes of accurate 
registry, the relative positions of the point on the two negatives should be 
the same. Let us divide the two negative materials into squares, and call 
the point where the axis of the lens meets the negative material, the origin 
of a co-ordinate system which locates the squares. Now the position which 
the point under consideration will occupy in each of the two negatives, will 
depend only upon the angle which the beam of light that originates at the 
point, will make with the axis of the lens. But if the two lenses have their 
axes separated from each other, as they must have if they are not situated 
in the same place, then the angle which the beams make with the two axes 
must necessarily be different. Hence the relative positions that the point 
will occupy in the two negatives will be necessarily different. Absolute registry 
will therefore be impossible under these conditions. The practical require- 
ments are so severe as regards registry, that even with a separation of a frac- 
tion of an inch in the optical axes of the system of lenses, the registry will be 
ruined and useless. But despite these insurmountable difficulties, inventors 
disclose cameras where the three separations are made through three lenses 
that are situated side by side, or in the form of a triangle, or in some other 
similar arrangement. It cannot be too strongly emphasized that correct 
registry under such conditions is impossible. Of course this is a great pity, 
since the use of three lenses would make possible the manufacture of cameras 
of great light efficiency, the only loss being that due to the absorption of the 
filters. 

The early inventors in this field include Pfenninger (Eng. P. 25,908/06); 
Christensen (Ger. P. 203,110; Eng. P. 7,514/08; U.S.P. 979,129); M. Maurich 
(Fr. P. 444,232; Ger. P. 264,085; Eng. P. 13,510/12); Featherstone (U.S.P. 
1,034,006), Gaumont (Fr. P. 437,173; Eng. P. 3,220/12), and Ulysee (Fr. P. 
459,669; Eng. P. 30,108/12). Later disclosures along these lines by G. Griffith 
(U.S.P. 1,589,754); Mannes and Godowsky of Kodachrome fame (U.S.P. 
1,619,949); R. S. Alldridge (U.S.P. 1,916,132; Eng. P. 322,801); C. L. Fitz 
(U.S.P. 1,931,983); and R. Thomas (U.S.P. 1,949,339). The disclosures of 
Thomas and Alldridge indicate some corrective measures to reduce the parallax 
caused by the non-identity of the optical axes, but at best these are of only 
doubtful value. Because of the inherent flaws, there is little need to continue 
the discussion of the various disclosures made in the above listed patents. 

A more serious attempt to compensate for the parallax due to the use of 



38 HISTORY OF COLOR PHOTOGRAPHY 

several lenses adjacent to each other has been disclosed by A. J. Arnulf in a 
patent (Eng. P. 435,222) which forms the entire Type 2. Here parallel plates 
of glass are placed in front of the lenses at definite angles, and the claim is made 
that the consequent shift in the position of the image points will compensate 
completely for the parallax. We shall have reason to discuss in Chapter 5 de- 
vices which contain parallel-faced reflectors and transmitters, so that it would 
not be amiss to follow in the footsteps of Major Klein and treat this dis- 
closure at some length. When light is incident upon a glass plate with par- 
allel sides, at angle i, it will emerge from the other side of the plate at the 
same angle, but it will be displaced a distance d which can be determined 
from the following expression 

d = E(i — i/n) sin i 

This expression holds when the angle i is less than ten degrees. The corre- 
sponding displacement 5 on the negative for thickness E, will be given by 

5 = df/D = Ef/D(i - i/n) sin i 

Here D is the distance that the point is from the objective, and/ is the focal 
length. When D is very large, that is, when the point being considered is 
at infinity, the displacement upon the negative will be zero. It becomes of 
considerable moment when the point is relatively near the objective. 

Consider now the case depicted in Fig. 15, where the point A is photographed 
by means of two objectives, and O2. Let us consider the objectives to be 
ideal ones, and let and O2 be the nodal points. The rays that come from 
the point A and pass through and O2 will continue without deviation to 
the focal planes of the two objectives. The intersections will be the image 
points in the two cases. Thus the image of A will be At when photographed 
by the first lens, and it will be Ai when photographed by the second objective. 
The points B and C, situated directly behind the point A when the observation 
is made through the lens 0, will yield images which will fall directly upon the 
image of A. But when observed through the lens O2 these points will no longer 
be seen as positioned directly behind the point A, but somewhat on the side 
as well as behind. The lens O2, will image these as three distinct points. 
This is the true meaning of parallax. There is no difference in image size, 
only a difference in the relative image position of the same point when photo- 
graphed through two lenses. If the point A be moved back to infinity, the 
lens would image it at Ai, but the lens O2 will image it at the point Z 2 . 
The distance ^2^2 is the extent of the parallax, and it is this amount that 
Arnulf attempts to correct for, by the use of a parallel plate glass placed in 
front of the two lenses. It is a very easy matter to compute the constants 
for a glass that will displace the image of the point A a distance equal to 
-42^2, but this will not be sufficient to correct for the parallax. What is re- 
quired is a scheme which will so distort the optical system of the lens 2 , 
that the images of the points A> B, C, D> E, etc., will all coincide, just as they 



COLOR CAMERAS 



39 



Qt 



*,$,T, A, 



FIG. 15 



do when the photography is done through the lens O. Not only must it com- 
press the images from the points A, B y C, etc., into a single image, but it 
must also do the exact opposite with the points R, S, T y etc., which lie upon 
the ray that emanates from Z off in infinity, and which is imaged at the 
point Z 2 by the lens O2 when no compensating devices are present. In the 
lens system 0, these will give rise to the distinct image points Ri, Si, 2\, etc. 
To escape from parallax, these should give rise to distinct image points in the 
second lens system. Thus the same optical device must compress the images of 
the points A, B,C, etc., into a single image, and it must spread out the image 
of the points R, S, T y etc., into distinct images. No effort is made in the patent 



40 HISTORY OF COLOR PHOTOGRAPHY 

disclosure to indicate that such would happen, nor is it indicated that such 
is even possible; all that is indicated is that it is possible to shift the relative 
position of the image any desired amount by placing a glass plate with*parallel 
sides at an angle in front of the lenses. All the corrective measures that were 
disclosed in the earlier patents suffer from the same cause. They shift the 
optical axes of the different lenses closer together, but they fail to correct for 
true parallax. True parallax means that points which give rise to a single 
image in one lens system, give rise to distinct and separate images when 
viewed by means of another lens system. 



CHAPTER 5 
COLOR CAMERAS, TYPE 3 



L 



lN the Type 3 classification, there is included those cameras where the light- 
splitting devices are situated behind the lens. The lens beam is divided into 
two or more sub-beams, the intensity of each of which is a fraction of that of 
the original. This type of camera is characterized by the fact that the sub- 
beams are all brought to a focus on different emulsions, but the points of foci 
all occupy the same relative positions in the three negatives. Thus there is 
completely eliminated the problems of parallax, provided the light-splitting 
mechanism itself does not introduce distortions. If due regard be taken of 
the optical requirements and properties of glass, such distortions need not 
appear. 

The simplest type of light-splitting device is represented by a half-silvered 
mirror. We will restrict the use of this term to designate reflecting surfaces 
that are uniformly coated with a thin deposit of metal, the deposit being so 
thin that it is partially transmitting and partially reflecting. It is possible to 
construct another type of transmitting and reflecting mirror where a glass 
surface is only partially silvered and partially free from silver. The silvered 
portions of the surface are complete reflectors, while the non-silvered portions 
are completely transmitting. We will call this type of reflector a partially 
silvered mirror. While we speak of half-silvered or partially silvered mirrors, 
it must be understood that the coating need not be restricted to silver, but can 
also be made of gold, platinum, chromium, aluminum, etc. 

The first camera disclosure was made by du Hauron in 1862. Here was 
disclosed the basic construction for a camera which utilized two semi-reflecting 
and transmitting mirrors, which divided the lens beam into three parts. This 
is illustrated schematically in Fig. 16. M± and Mi are transparent mirrors, 
while Mz is a completely reflecting mirror. is the optical center of the lens 
system. The dotted lines represent the path traversed by the light rays. 
Let us consider the axial ray. This strikes the first mirror at b where it is 
partly reflected and partly transmitted. The reflected beam is brought to a 
focus in the first image plane I\ at the point p. The transmitted beam is 
again partly reflected and partly transmitted by the second mirror M 2 . The 
reflected part is brought to a focus in the image plane h, giving rise to the 
point /, while the transmitted beam is directed to the mirror Mz from which 
it is brought to a focus in the third image plane h at the point i. The three 
points p> tj and i all occupy the same relative positions in the three planes, 

41 



42 



HISTORY OF COLOR PHOTOGRAPHY 



image, plane h 




Ma 
ftitrror 



1 




1 
1 


1 \ 


1 
1 


1 \ 

1 \ 

1 I 


1 


1 \ 


1 
/ 


1 \ 

I 



Ms 

rofjectutg' 



FIG. 16 



so that were they superimposed upon each other, the three points would fall 
one directly upon the other. Thus under these conditions no parallax is 
possible, unless the mirrors introduce distortions. 

A modification of this type of camera was disclosed by A. H. Cros in an 
English patent, No. 9012/89. Here was disclosed for the first time the use 
of a rotating mirror Mi (Fig. 17) which revolved about its axis rs. Parallel to 
the plane of this mirror was a stationary mirror M2. The rotating mirror 
was silvered only at certain sections, these being so arranged that the mirror 
transmitted light directly to the image plane 7i part of the time, and the rest 
of the time it reflected the light to the mirror M 2 . This reflected the light back 
again to the mirror ikfi, but to a different segment of it, and the silvered and 
non-silvered sections of Mi were so arranged that part of the time the beam 
that is reflected from the near end of M 2 will pass through a non-silvered 
portion of Mi, and the rest of the time it will be reflected again by Mi to the 
far end of M2. From here it is reflected to the image plane J3. The transmitted 
beam is imaged at Z 2 . Thus the same lens beam gives rise to three image points 



COLOR CAMERAS, TYPE 3 



43 



in three different planes, which occupy the same relative positions in the 
respective negatives. 

The original du Hauron chromoscope described above contains the really 
basic elements of camera design. In other patents, these elements were further 
developed. Thus in the original, du Hauron used silvered mirrors for his 
semi-transparent reflectors. In a later patent issued in 1874 (Fr. P. 105,881; 
Eng. P. 2,973/76) he disclosed another new idea. Instead of using a half- 
silvered mirror, one which is transmitting because of the thinness of the silver 




FIG. 17 

deposit, he stated that it would be possible to achieve the same result if a 
mirror which was completely reflecting would be so treated that certain areas 
would have the silvering completely removed. These portions would then 
become completely transmitting, and the mirror would be composed of al- 
ternate areas that were completely transmitting and completely reflecting. 
The ratio of the transmission to the reflection could be governed by varying 
the respective areas of the silvered and non-silvered portions. The light 
efficiency of half-silvered mirrors is far from good. Thus, a mirror which 
has a surface coating of silver sufficiently thick to transmit twice as efficiently 
as it reflects will absorb about 30 per cent of the light that is incident upon it. 
This light is absorbed by the metallic deposit. If the deposit is made sufficient 
so that the mirror is completely reflective, then the amount of light that is 
absorbed by the metallic deposit is very low, not more than 5 per cent. Hence 
the du Hauron innovation gave a considerable increase in light efficiency. 

The use of reflecting surfaces that are silvered only in selected areas has 
proven to be quite popular, and other inventors included such surfaces or 



44 



HISTORY OF COLOR PHOTOGRAPHY 



modified forms of them in many successful cameras. Ives, for instance, a 
prolific inventor in the entire field of color reproduction, designed a camera 
where prisms instead of glass reflectors were used to split the light. The prism 
surface that actually divided the beam consisted of alternate bands of silvered 
and non-silvered portions (U.S.P. 703,929). A different form of the same 
idea is found in the French patent 350,004, issued to the Lumieres. The 
beam divider consisted of two sheets of glass that had alternating silver and 
clear areas upon them. In one of the glasses the stripes ran horizontally, and 
in the other they ran vertically. The two sheets of glass were placed at right 




FIG. 18 



angles to each other, forming a cross. The silver stripes were parallel to the 
intersection in one of the glasses, and perpendicular to it in the other (Fig. 18). 

The idea crops up again in several of the very early Technicolor patents. 
In the English patent 101,972, and again in the United States patent 1,231,710, 
Technicolor disclosed light-splitting devices which utilized surfaces on which 
was spluttered metal in such a fashion that irregular patches of totally re- 
flecting polygons were formed. This created a grid that consisted of minute 
areas that were totally reflecting, separated from each other by areas which 
were totally transmitting. This is a real improvement over the others since 
it removed any unevennesses that might be caused by the banded type of re- 
flectors. A slight modification of this is disclosed in United States patent 
1,451,774, issued to Holbrook and de Francisco. These gentlemen prepared the 
reflecting surface in the shape of a checkerboard, so that each transmitting 
area is surrounded by similar areas which are completely reflecting. 

The disclosures discussed above all utilized glass surfaces or glass prisms 
to carry the metallic deposits. Where a glass surface is used, there is con- 



COLOR CAMERAS, TYPE 3 45 

siderable danger that distortion is introduced. This is caused by the refraction 
of the rays as they penetrate the thickness of the glass. This is best illustrated 
in Fig. 19. Here represents the optical center of a lens system, which would 
normally give an image in the focal plane of the lens. Now let us place a 
reflector in the path of the rays from 0, at an angle of 45. degrees. Consider the 
axial ray Ob. In the absence of the reflector, this ray would give rise to the 
image point m, and the distance Om would be equal to the- focal length of the 
lens. But when the glass reflector is present, the ray impinges the reflector at 
b, and instead of continuing on in a straight line to s, it becomes refracted and 
emerges at the point e, after which it continues along a line parallel to the 






Cy 


7 


_— — 


1 




'e 






1 


Cc^: 


^^-^ 


cl^. 


1 
1 
1 



FIG. 19 

direction Os. The optical path which this ray traverses must be identical to 
that of the normal ray. Let us suppose that the focal length of the lens is 
p inches. Then Om is equal to p inches, since the index of refraction of air 
is 1. 000. The index of refraction of glass is 1.500. The optical path of the re- 
fracted beam is then equal to Ob + be/ 1.500 + a distance along eh sufficient 
to make the sum total to p inches. Therefore the focal plane becomes moved 
back a trifle farther away from the lens. Not only is the focal plane moved 
back, but the axis of the lens drops a distance equal to mh'. This distance de- 
pends upon the angle which the ray makes with the reflector, and a glance 
at the diagram will quickly demonstrate that the terminal rays will make 
angles with the reflector different from that of the axial ray, and that all of 
them will be different from each other. Therefore gg' the distance which 
the upper terminal ray Ocfg is depressed will be different from mh f , the distance 
that the axial ray is depressed, and that is different from nk', the amount 
that the lower terminal ray is depressed. If the thickness of the reflector is 
0.10 inch, then the difference between gg f and nk' may be several thousandths 
of an inch, an amount that would cause serious registry trouble where the 



46 HISTORY OF COLOR PHOTOGRAPHY 

negatives are to be enlarged to any size at all. This defect was very quickly dis- 
covered by the various inventors, and many steps were taken to correct for it. 

Ives (U.S.P. 622,480), and Strass Collins (Ger. P. 102,206) suggested the 
use of wedge-shaped glasses, the thicker ends of the glasses being placed in the 
path of the rays which make the smallest angle with the glass. Conrady and 
Hamburger (Eng. P. 28,722/12) would compensate for this distortion by 
tilting the focal plane somewhat. This has also been patented by W. Berm- 
pohl (U.S.P. 1,951,896) in 1934, fully a generation later. Probably the best 
solution for this problem was offered by O. Pfenninger (Eng. P. 25,907/06). 
He placed a similar plate of glass in the optical path, but at exactly the opposite 
angle. Thus an equal but opposite distortion was introduced, and one negated 
the other. This distortion is completely absent when prism blocks are used 
instead of glass plates, but their expense does not make their use very popular. 
Besides the expense, the use of a prism block necessitates the use of special 
lenses, for the normal lens is corrected only for the glass distance that the rays 
travel through the lens. This distance would be multiplied by at least two in 
the case of prism blocks. The Pfenninger solution, while cheap and of little 
consequence upon the normal lens corrections, makes the camera somewhat 
bulky. It is not always convenient to place reflectors and compensators in- 
side the limited space within the camera box. 

A very ingenious scheme which offsets this defect was disclosed by P. D. 
Brewster. Instead of using a glass surface for the reflectors, he used highly 
polished metal surfaces, and he made them partly transmitting by the mere 
expedient of drilling holes, so that only selected areas of the metallic mirror 
acted as reflectors, the remaining portions being open and devoid of metal. 
Because of their construction these reflectors were termed "Swiss Cheese 
Mirrors" (U.S.P. 1,253,138; Fr. P. 483,761; Eng. P. 100,082). In one form 
of the disclosure, the perforations in the metallic mirror were a series of parallel 
bands, duplicating the original du Hauron, Ives, and Lumiere banded mirrors. 
In another form the perforations were in the form of small holes, whose sizes 
changed toward the edges of the mirror. This is a duplication of the Comstock 
disclosures. In practically all of his many subsequent patents, Mr. Brewster 
retained the idea of metallic mirrors rather than those of glass. There is one 
disclosure, however, in which Mr. Brewster designed a light-splitting device 
which used a glass plate surface coated with silver or platinum to make it 
semi-transparent. This plate was immersed in a cell which contained a liquid 
such as Nujol or other mineral oil, and which had an index of refraction that 
was the same as the glass. The cell sides were also made of glass with the 
same index of refraction. As is evident from the physics of this situation, 
the net result was the same as if a prism-splitter were being used. This little 
scheme made the cost of the light splitter a fraction of the cost of a similar 
glass prism. The defect of such a procedure lay in the fact that the lens 
corrections were completely negated. The use of prisms necessitates specially 
corrected lenses. This has been disclosed in United States patent 1,277,040. 



COLOR CAMERAS, TYPE 3 47 

Another method whereby the evils of refraction are eliminated is by the 
use of pellicle reflectors. A camera where this is used is disclosed by Reck- 
meier (U.S.P. 1,895,555; Eng. P. 357,37 2 )- Here only one of the reflectors 
is of this type. The pellicle reflector is characterized by its extreme thinness, 
approximately 0.005 mcn - It is usually made of cellulose, celluloid, or cello- 
phane that has been stretched taut in a frame, so that the pellicle will lie 
evenly and without distortion. When they were first introduced, it was not 
possible to coat them with metallic reflecting surfaces, but there have lately 
been developed methods of metallic spluttering (Eng. P. 440,006) which give 
the proper metallic semi-transparent mirror surfaces. Methods of casting 
such pellicles have been disclosed by Adrian Klein (Eng. P. 380,938). The dis- 
placement of the image due to a pellicle that is 0.005 inch thick is completely 
negligible. 

The use of crossed mirrors has been disclosed in quite a few patents. Some 
American patents where such mirrors are used are the following: 1,217,391 to 
C. N. Bennett; 1,631,058 to H. Piloty; 1,812,765 to M. Astafiev; 1,857,578 
to W. L. Wright; 1,897,097 to W. L. Wright; and 1,986,425 to O. O. Ceccarini. 

In camera systems where semi-transparent mirrors are used, each negative 
receives the light from the full beam that enters the lens, but at a greatly re- 
duced intensity. This was not the case with the camera designed by A. H. 
Cros and described above. As will be recalled, Mr. Cros used a rotating 
mirror to deflect the full lens beam to another mirror, which then deflected 
it to the proper negatives. This camera was disclosed in 1889, a full generation 
before Mr. Brewster, who has sometimes been credited with the first intro- 
duction of rotating mirrors. This is rather difficult to see since there have 
been several other predecessors. Dallmeyer (Eng. P. 22,616/98) proposed to 
use a camera in which a single mirror was placed behind the lens, the mirror 
then being rotated to reflect the light to two sides of the camera. A similar 
idea was used by Gaumont (Jahrbuch, Vol. 16 (1902), p. 543). In 1919 
J. Dourlen and M. Chretien (Eng. P. 141,368) disclosed a camera which used 
rotating mirrors to deflect the lens beam to three sides of a rectangle, but this 
had evidently been previously disclosed by Gebay in 1913 (Fr. P. 464,637). 
Mr. Brewster's disclosure did not appear until 1930 (U.S.P. 1,752,477 and 
1,860,525). In an article published in the Journal of the Society of Motion 
Picture Engineers (Vol. 16, p. 49), Mr. Brewster states the case for rotating 
mirrors. It has the advantage of not having to transmit the image through 
a block of glass, which results not only in a loss of light, but also in a loss of 
definition around the edges. But most important of all it makes possible the 
use of lenses with a focal length as little as 50 mm at an aperture of/: 2, allow- 
ing the placement of the lights fairly close to the object. This gives better 
efficiency for the light. Just slightly later than Mr. Brewster, J. A. Ball of 
Technicolor Motion Picture Corporation also disclosed the use of revolving 
mirrors (U.S.P. 1,871,649 and 1,924,901). 

An entirely different principle for dividing the lens beam into two or more 



48 



HISTORY OF COLOR PHOTOGRAPHY 



sub-beams is disclosed by Ives in United States patent 632,573 (1899). In 
this disclosure the cross section of the lens beam is divided into three parts 
by means of prisms which lie immediately behind the lens. Thus (Fig. 20) 
the pencil of parallel rays which enters the lens comes from a common point 
on the object being photographed. Immediately behind the lens there are 
placed two prisms. Part of the lens beam, marked a in the diagram, is incident 
upon the upper prism. These rays are deflected to the point/. Another part 




FIG. 20 



of the lens beam designated by the letter c, is incident upon the lower prism, 
and these rays are deflected to the point g. The central portion of the beam 
passes through an opening between the prisms, and this is brought to a focus 
at the point A. In order to equalize the optical paths of the three beams, a 
glass block is placed in the path of the central rays which do not pass through 
any of the prisms. 

Five years later, in 1904, O'Donnell & Smith (Eng. P. 4127/04) devised a 
camera where the same result is obtained by the use of mirrors directly behind 
the lens. This is illustrated in Fig. 21. Here again the lens beam abc emanates 
from a single point on the object being photographed, hence the rays are 
parallel to each other. That portion of the rays marked a is reflected by a 
total reflecting mirror M to the point d. Another portion passes through 
directly to the point e. The third portion marked c, is deflected by the mirror 
N to the point /. Hence a common point on the object becomes imaged at 
three different points, d, e, and /, all occupying the same relative positions 
in the three focal planes Pi, P 2 , and P 3 . Much more recently, the Jos Pe 
Company disclosed the same idea in several of its patents (U.S.P. 1,597,818; 



COLOR CAMERAS, TYPE 3 



49 




FIG. 21 

Eng. P. 243,714 and 243,716 issued to H. Piloty; U.S.P. 1,801, i43|issued|to 
R. F. P. Defregger). 

The Ives disclosure was somewhat changed in a patent issued to Dr. L. T. 
Troland of Technicolor (U.S.P. 1,843,007; Eng. P. 350,112). Here a prism 
block is placed immediately behind the lens system (Fig. 22). The prism 
block is designated by the shaded area in the diagram. The surfaces M and 




FIG. 22 



5° 



HISTORY OF COLOR PHOTOGRAPHY 



N are semitransparent. Therefore the image in the plane P 2 is constructed 
from the entire lens beam. The upper portion of the beam, which is incident 
upon the surface M of the prism, is partly reflected to the image plane Pi, 
and there gives rise to the image point d. The image point/ is similarly formed 
from the portion of the lens beam that is incident upon the surface N. Funda- 
mentally, this does not differ from the original Ives disclosure, except in the 
details of construction. This disclosure has one great advantage over that of 
Ives in that the path of the rays through glass is identical in all three cases, 
so that there is no need to place corrective glass blocks in the path. Dif- 
ferent only in structural details is the camera designed by W. M. Thomas 
(U.S.P. i,9S7,37i and 1,988,882). 



a — 




h 



The simplest and most popular types of cameras go back to the original 
design of du Hauron (cf. Fig. 16). Here the beam that enters the lens is 
made incident upon three reflectors one after the other, and each reflector 
deflects part of the beam to a different image plane. Edwards (Eng. P. 
3,560/99) reduced the number of reflectors to two. He also disclosed the use 
of colored reflectors, thus eliminating the presence of double images. Part 
of the ray "a" which enters the lens (Fig. 23), is deflected by the top surface 
of the mirror Mi to the image plane Pi, where it is brought to a focus at the 
point /. The remainder of the ray continues to the second deflector M2. 
But the bottom surface of the mirror M\ also acts as a reflector, and this would 
give rise to another image point in the plane Pi, which would not coincide 
with the point / since this image traversed two thicknesses of glass at the 
point b. But if the reflector M\ be colored with a substance which absorbs 
the wavelengths that are to be registered in the image plane Pi, then this 



COLOR CAMERAS, TYPE 3 



5i 



second image point would not be registered. Let us suppose that at Pi we 
register the blue primary. Therefore in front of the emulsion, or somewhere 
in the path bf, there is placed a blue filter. Now if the mirror Mi be colored 
yellow, the ray that will be reflected from the bottom of this mirror will 
be completely lacking in blue light, hence will not be able to register in the 
plane Pi. The ray bf, which is reflected by the front surface of the mirror will 
not be affected by the coloring matter. If the image plane P 2 is made to register 
the red primary, then the mirror Mi should be colored green, so that no red 
rays will be reflected from the bottom surface of that mirror. • 

It is possible to arrange the mirrors in many different ways in order to 
obtain three images by this system. This has given rise to a large number 
o£ patents, each utilizing the same fundamental principles, but each having 
a more or less complicated arrangement of the mirrors. Nachet (1895); 
White (1896); Butler (1897); and Dawson (1912), were some of the other 
early inventors in this field. Somewhat later came Albert (U.S.P. 1,607,661); 
G. Cuenin (U.S.P. 1,637,294); Bermpohl (U.S.P. 1,951,896); Jacobsen (U.S.P. 
2,082,579); and many others too numerous to mention. 

Everything that can be done by means of mirrors, can also be accomplished 
by means of prisms, and every single disclosure which utilized a mirror has its 
counterpart in a patent which does the same thing by the use of prisms. 



*_!- ap_ -j 




FIG. 24 

Technicolor, outside of the few patents that concern themselves with metallic 
rotating mirrors, seems to have confined itself solely to the use of prisms for 
light-splitting devices. Thus Comstock and Ball disclosed a rather ingenious 
device which gave two images upon adjacent frames of a film (U.S.P. 1,451,325; 
Fig. 24). The prism block is in the form of a triangle. It is really composed 
of two right-angled triangular prisms cemented together along one side. 
This common juncture between the two prism components, marked M in 
the diagram, is semi-transparent. The lens beam entering the prism block is 
therefore divided into two beams when it strikes this surface. One of the 
beams continues on to the surface N 2 , where it is totally reflected to the image 
plane P 2 , while the other beam is reflected back to the surface N x from which 
it is again reflected to the image plane P x . The application to three-color 



52 



HISTORY OF COLOR PHOTOGRAPHY 



separation is of course obvious. The same idea crops up again in a patent 
issued to Dr. Troland (U.S.P. 1,821,680). 

There may be some advantage if the three images can be placed upon the 
same piece of film, or upon the same plate. This means that the optical de- 
vice must be such that the three image planes fall in a common plane. One 
of the very first to do this was Ives (U.S.P. 660,442) in 1900. The lens beam a 
(Fig. 25) first strikes the semi-transparent mirror Mi at the point b. Here 



a — 



ci 






\k -J/ 




h 

f 

h 

u 



FIG. 25 



it is divided into two beams. One of these continues to a second semi-trans- 
parent mirror M 2 where it is again divided into two beams. The total division 
is therefore into three beams. One of the beams is reflected off at d to the 
prism block A where it is totally reflected at the point e to the image plane I\. 
Another beam continues from d in a straight line to the image plane I 2 . The 
third beam starts at the point b and by means of a reflection by a mirror Ni, 
is directed to the image plane 7 3 . A glass block B is interposed in the path 
of this beam. The thicknesses of the glass blocks A and B are such that they 
equalize the three optical paths, and allow the three image planes to exist 
in a common plane. The three images could therefore be imaged upon the 
same piece of film or plate. Selle, in an early English patent (Eng. P. 12,514/99), 
used a similar system of mirrors to accomplish the tripartite division of the 
beam, but instead of using glass blocks to equalize the optical paths of the 
two outer beams, he put an auxiliary lens into the central beam to shift its 
focal plane forward until it lay in the same plane as the other two. Kunz 
(U.S.P. 1,319,292) combined the mirrors with the glass blocks so that a single 
block accomplished the same result. His scheme is indicated in Fig. 26, 
which is self-explanatory. Isensee (Ger. P. 334,776) accomplished a similar 
result, but instead of using a prism block made of glass, used a cell whose 
sides were glass, but whose interior was filled with cinnamic ether, which 
has a very high index of refraction. This is an obvious extension of the 
Brewster idea previously discussed. 



COLOR CAMERAS, TYPE 3 



53 



d 




FIG. 27 

A somewhat different arrangement was disclosed by L. Albert (Fr. P. 
554,056; U.S.P. 1,607,661). Six mirrors were used. Of these (Fig. 27) two 
were semitransparent, M\ and Mi, while the others were totally reflecting. 
The lens beam a, after reflections from the six mirrors, gave rise to the three 
image points h, j, and k all in the plane /. 



CHAPTER 6 



COLOR CAMERAS, TYPES 4 AND 5 



I 



N a previous chapter it was pointed out that the cone of light which starts 
from any point on the object and enters the camera lens can, to all intents 
and purposes, be considered as a pencil of parallel rays. The position on 
the negative where these rays are brought to a focus is determined solely by 
the angle which the beam makes with the optical axis of the lens. If we have 
a system that utilizes two or more lenses, then it is incumbent upon the de- 
signer to see that the various sub-beams make identical angles with the optical 
axes of the separate lenses. It is only under those conditions that a given 
point will occupy the same relative position on all negatives. 

Now let us consider a beam of light which emanates from a single point 
on an object, and which is split up into three beams by a system of semi- 
transparent mirrors. Each of the three beams will be made to enter a lens. 




L/ 



* 



La. 

f 



i 






FIG. 28 

This type of system, where the light-splitting device lies in front of the lens 
system, characterizes color cameras grouped under Type 4. A very simple 
schemealong this linewas disclosed by Procoudin-Gorsky (Eng. P. 185,161/15). 
The beam of light, s, emanating from the object being photographed (Fig. 28) 
is first incident upon a total reflecting mirror 0. The beam continues in the 
new direction at right angles to the old, until it strikes a semi-transparent 
mirror b. Part of the beam is again reflected at right angles, and is brought 
to a focus at the image plane I\ by the lens L\. The rest of the beam is trans- 

54 



COLOR CAMERAS, TYPES 4 AND 5 



55 



mitted to a second semi-transparent mirror c, where the same thing occurs. 
The reflected beam is brought to a focus in the image plane h by the lens Z2. 
The transmitted part is totally reflected by the mirror d, and it is then brought 
to a focus by the lens L 3 in the image plane 7 3 . A glance at the diagram shows 
quickly that the three sub-beams make identical angles with the optical axes 
of the three lenses, hence the points e, /, and g all occupy the same relative 
positions in the three negatives. 

Because this scheme is the first mentioned, it must not be inferred that 
it is the first mechanism disclosed which utilized a light-splitting device in 
front of the lenses. A similar type of technique was disclosed in 1900 by that 




FIG. 29 

prolific inventor in color photography, Ives (U.S.P. 660,442). The beam 
coming from the object strikes a semitransparent mirror (Fig. 29) M 3 , and 
is partly reflected to the second semitransparent mirror M* which is at right 
angles to the first. The first part beam is reflected at a to the mirror M h 
which in turn reflects it to the lens Li. The second part beam is again partly 
reflected at c to the mirror M2, which reflects it in turn to the lens L%. The 
transmitted beam at c is directed to the lens L2. It is quite apparent from the 
diagram that the paths of the three beams are not the same, but that the two 
reflected beams travel somewhat greater distances than the direct beam. In 
order to equalize the three paths, Ives inserted blocks of glass, P, in the paths 
of the reflected beams. 

The insertion of the blocks of glass created some problems, since they had 
a tendency to obliterate the lens corrections, which were calculated only for 
the glass distance in the lens system itself. Sacre (Fr. P. 460,310) disclosed 
practically an identical scheme, except that he left out the glass blocks which 
equalized the optical paths. This type of camera gave a slight parallax for 
the central beam, but if this is used for the blue filter image, it may not be 
very noticeable. Both Ives and Sacre place the three images upon the same 
negative, that is, in the same plane. This may be desirable from the point 
of view of convenience of operation; Horst (D.G.M. 757,468) accomplished 



56 



HISTORY OF COLOR PHOTOGRAPHY 



this in a rather simple fashion, by merely by-passing the beams given distances, 
until the optical paths became identical. The diagram illustrating this (Fig. 30) 
is self-explanatory. M and N are total reflecting mirrors, R and S are semi- 
transparent mirrors. 

1 
y 



1, 




■0---i^ : /K 




FIG. 30 

A more complicated scheme is that of W. L. Wright and S. MacDonald 
(U.S.P. 1,641,466), and W. L. Wright (U.S.P. 1,688,606). In the second of 
these, the beam is made incident upon a series of semitransparent and totally 
reflecting mirrors (Fig. 31) which divide it into three sub-beams, each of which 

Ma 

I: 




^ 



4 



L.3 



FIG. 31 

is directed to a different lens. The three images lie in the same plane. A 
somewhat different scheme is disclosed by W. L. Wright in two patents 
(U.S.P. 1,730,712 and 1,930,498). Here the beam meets three mirrors, one 
after the other, so arranged that three images are formed in the same plane 



COLOR CAMERAS, TYPES 4 AND 5 



57 



by means of one lens. The first semitransparent mirror, Mi (Fig. 32) meets 
the beam at an angle that is slightly less than 45 degrees to the lens axis. The 
second mirror, M2, is inclined at an angle of 45 degrees. The third mirror, 
M 3 , is inclined at an angle slightly greater than 45 degrees. The three beams 
deflected by the three mirrors meet at the front nodal point of the lens system, 
L, and are imaged by L at the three points d, e, and /, all lying in the same 
plane. To equalize the optical paths, compensators c\, C2, and c% are placed 
between each mirror and the lens. These are blocks of glass of sufficient 
thicknesses to make up for the difference in the beam lengths. This means 




FIG. 32 



that the distance of must be equivalent to abe, and these two must be equiva- 
lent to acd. Compensation is brought about by the fact that the optical 
distance that a beam travels through glass is equal to the thickness of the 
glass divided by its index of refraction, which is usually about 1.5000. There- 
fore if a beam travels for a distance of one and one-half inches through glass, 
it is equivalent to a distance of only one inch in air, which has an index of 
refraction of 1.000. Therefore the block c 3 must be much thicker than the 
block c 2 , and this in turn must be thicker than the block C\. This scheme 
may equalize the optical paths of the three beams, so that the points d, e, 
and/ will fall upon the same plane, but it does not guarantee that the images 
due to beams at other angles will occupy the same relative positions with 
respect to each other, in all three images. 

Most of the inventors in this field seem to have gone to great lengths to 
place the three images in the same focal plane, so that all three images could 
be photographed at different portions of the same plate. In certain respects 



58 



HISTORY OF COLOR PHOTOGRAPHY 



this may seem like a good idea, in that but a single plateholder is required in 
the case of still pictures, and a single film-moving mechanism in the case of 
motion pictures. But this insistence leads to complications in the optics of 
what would otherwise be a very simple system. F. A. Bourges discards the 
idea completely, and very frankly couples three cameras behind a mirror 
light-splitter of the type disclosed by Procoudin-Gorsky. The three camera 
movements operate in unison, thus insuring registry. This is disclosed in an 
English patent (460,335). Another disclosure where three separate image 
planes are used was patented by R. E. Reason (U.S.P. 2,053,224; Eng. P. 
418,671 and 423,190). Here there is a peculiar combination of types. The 
main beam (Fig. 33) is split into two by the semitransparent mirror Mi. 




FIG. 33 

The reflected beam is imaged by lens Zi in the image plane h. The trans- 
mitted beam passes through lens Z2, then is split into two parts by the mirror 
Ms, giving rise to images in the planes h and J 3 . It is quite evident that re- 
gardless of the angularity which the beam makes with the optical axis of the 
lens Li, it makes the same angle with the axis of the lens L$. 

A. G. Hillman, in a series of patents (Eng. P. 404,307, 414,059, 414,065, 
and 434,719) discarded the idea of equalized light paths. The light beam 
was split into two by means of a system of perforated and unperforated 
metallic mirrors, and each beam was deflected into a separate lens. The 
two lenses were situated one immediately above the other. Obviously the 
beam which entered the upper lens traveled a greater distance. The film 
(this idea was disclosed for use in motion pictures only), traveled past the 
two lenses one frame at a time. Two adjacent frames were exposed at one 
time. After each exposure, the film advanced one frame, and at the same 
time the filters changed, so that the filter which was in front of the top gate 
shifted to the bottom one. Let us suppose that at a given exposure, the red 
filter is on top and the green filter is on the bottom. In this manner, the 



COLOR CAMERAS, TYPES 4 AND 5 



59 



top frame is exposed to the red primary, and the bottom frame to the green. 
After the exposure, the film advances one frame while the filters change posi- 
tions. For the second exposure, the top frame is the same one which had 
previously received an exposure behind a green filter. In its new position, 
since the filters also changed, it is again behind the green filter. Thus each 
frame receives two exposures through the same filter. The film advances 
through the camera at the standard black-and-white rate. The separations, 
two-color, are present as alternate frames in the finished negative. It is 
claimed that because each frame receives two exposures, the parallax due to 
obvious inequality of optical paths is ironed out, and therefore not noticeable. 
This scheme can be used for three-color also, either by the use of three lenses, 
or by the use of two lenses but three filters. 

Another rather ingenious scheme utilizes a single lens behind a mirror system, 
to give two images. This was disclosed by Bennett (Eng. P. 10,150/12) 
considerably before Wright. Here (Fig. 34) the beam of light falls upon a 




FIG. 34 



totally reflecting mirror Mi situated above the lens. This reflects the beam 
to two mirrors situated in front of the lens and at a slight angle to each other. 
The lens will image the beams from the two mirrors M2 and M z at different 
portions of the same plane. By having the angle between the two mirrors 
at a proper value, it is possible to maintain the equality of the optical paths 
and also the equality of the angle which these beams make with the optical 
axis of the lens. In this case, the lens does not directly image the object being 
photographed, but images the reflection of the object in the mirror Mi. 

Everything that can be done by mirrors can also be duplicated by prisms. 
The patent literature is fairly replete with disclosures of various prism designs 
which split the light into two or three beams. J. F. Romer, in a series of 
patents disclosed numerous designs of prisms which receive a single beam of 
light, and which transmit two, three, and four beams to an equivalent number 
of lenses. It is therefore possible to image two, three, or four images upon 
the same plate (U.S.P. 1,722,356, 1,722,357, and 1,765,882; Eng. P. 257,546, 



6o 



HISTORY OF COLOR PHOTOGRAPHY 



257,547, 257,548, and 257,549). Twyman (Eng. P. 16,811/15; cf. also, 
16,810/15) placed a simple prism block in front of the two lenses (Fig. 35), 
and a compensating block behind the lens which received the reflected beam. 
Practically the identical scheme, changed only to place another compensating 




FIG. 35 

glass block in the path of the transmitted beam, was patented considerably 
later by V. Hudeley (Eng. P. 444,051). 

K. Martin and P. Tietze placed a simple prism block (U.S.P. 1,752,680; 
Eng. P. 276,591) in front of two lenses. Behind these lenses were deflecting 
prisms, designed to direct the two beams so that they emerge as parallel 
beams and are brought to a focus in the same plane (Fig. 36). Other prism 




FIG. 36 



devices were disclosed by A. Schustek (U.S.P. 1,845,062); D. Daponte 
(U.S.P. 1,945,029; Eng. P. 346,406); C. LeRoy Treleaven (U.S.P. 1,696,739); 
T. A. and R. T. Killman (U.S.P. 2,060,505); and A. Pilny (Eng. P. 342,036). 
Obviously there are infinite numbers of variations possible. F. E. Tuttle, 
of the Eastman Kodak Company, disclosed a novel method to equalize the 
optical paths of the different rays, by shifting the nodal points in the lenses. 
This is disclosed in U.S.P. 1,897,874. A simple prism divides the beam into 
two sub-beams, and directs them to two lenses. These lenses differ from each 



COLOR CAMERAS, TYPES 4 AND 5 



61 




FIG. 37 

other in that the front nodal point F on one lens is shifted a distance equivalent 
to the difference in the optical paths of the two rays (Fig. 37). 

L. Didier placed a simple prism system in front of three lenses. Only two 
of the lenses (Fig. 38) received the light from the same beam, while the third 
lens received the light from a different beam. No effort was made to equalize 
the optical paths of the beams. Parallax was therefore present on two ac- 
counts (Eng. P. 306,329). Similar schemes were disclosed by J. M. Gutmann 
and P. Angenieux (Eng. P. 437,414, cf. also 385,141 and 419,894). 



«<»- 



s 2 > 



f 



i/ 



4 



Lz 



4 



U 



FIG. 38 

A series of very important patents were issued to J. H. Dowell and Adam 
Hilger, Ltd. In these patents (U.S.P. 1,839,955; Eng. P. 349,107, 3^8,754, 
and 427,983) there is discussed the fundamental requirements for a prism 
system which will give two or more images that are identical in size and free 
from parallax, after transmission through a colored filter. Such prisms are 
constructed of glass, and it is usually found that the identity of sizes formed 
through these prisms is satisfied only for objects at infinity. When objects 
are not at infinity, it is found that the identity requirement is no longer 
satisfied. 

In order that this be true for all objects, both near the lens and at infinity, 
two conditions must be satisfied. The first is the usual condition which prac- 



62 



HISTORY OF COLOR PHOTOGRAPHY 



tically all inventors satisfy, the condition that the reduced optical paths of 
the beams and sub-beams be identical. Thus if a given beam travels through 
air and glass, the following relationship must hold for each color. This is 
illustrated in Fig. 39. The beam enters the prism at a and proceeds to the 
semi-transparent mirror M\. Here part of the beam is reflected to the totally 
reflecting mirror M 2 , while the rest of the beam is transmitted to the mirror 
M3, where it is reflected to the mirror M 4 . The one beam travels through 
glass a distance equal to ab + be + cd, then through air a distance dr, where 




FIG. 39 



r is the front nodal point of- the lens. The total reduced optical path which 
this beam traverses will be, supposing that this beam registers the green 
image, 

(ab + bd + de) .dr 



N 



1. 000 



where N g is the index of refraction of the green rays for the glass used. The 
index of refraction of air is taken as 1.000. The other beam will traverse a 
glass path equal to (ab + be + ef) and an air distance equal to fs, where 5 
is the front nodal point of the second lens. This beam registers the red 
primary, so that the optical path traversed by this beam will be given by 



(ab + be + ef) 

N r 



fs 



1. 000 



where N r is the index of refraction for the red rays. The first condition to 
be satisfied is then the following: 

(ab + be + cd) , dr (ab + be + ef) 



fs 



N 



I. OOO 



N r 



I. OOO 



The second condition that must be fulfilled deals with refraction. When- 
ever a beam of light enters a glass block at an angle, the beam is refracted, 
and not only is the whole beam shifted, but the amount of shift is different for 
each color. When the beam emerges from the prism, there will occur a lateral 
displacement of the beam, hence a lateral displacement of the image of that 



COLOR CAMERAS, TYPES 4 AND 5 



63 



beam. The amount of this displacement is proportional to the distance which 
the beam travels through the glass, and to the tangent of the angle of refrac- 
tion. Therefore, if L is the total glass distance which a given beam traverses, 
and r is the angle of refraction, then the second condition states that 

2(£ tan r) red = 2(L tan ^green 

It can now be seen that if the distances which the separate beams travel 
through glass are not identical, then the displacement of the image due to 
refraction will be different, and the object will no longer be imaged in relatively 
the same positions in all negatives. The first condition does not state that the 
distance through glass must be the same for all beams, for it is possible to 




FIG. 40 



compensate a distance through glass by some smaller distance through air. 
From the point of view of image size, it is the total optical distance which 
must be identical, and this condition, as has been stated before, is satisfied 
by most of the inventors. But very few inventors have taken the trouble 
to satisfy the second of these conditions, which most probably explains why 
practically all of the disclosures listed above have never left the laboratory 
stage. In their patents, Dowell and the Hilger company outline several 
schemes whereby these conditions are fulfilled. These are illustrated in 
Figs. 39 and 40. The patent disclosures give detailed calculations for these 
cases. 

This brings our discussion to the fifth, and final, method whereby a beam 
of light emanating from a single point on the object, is divided into two or 
more beams, and brought to a focus upon two or more separate films. This 
method is characterized by the fact that it contains no mirrors, the beam 
division being accomplished by means of the lens system. We are confronted 
with the fundamental problem that a pencil or cone of rays, starting from a 
single point on the object, must be imaged as a single point. Why is it not 
possible, therefore, to pass this cone of rays through a lens which will main- 
tain the cross section of the pencil of rays at a maximum, and then interpose 
a series of small lenses across that cross section, and grind these lenses so 
that they will bring the section of the beam which they encompass, to a point 
focus? Then it would be possible to divide the original beam into as many 



64 HISTORY OF COLOR PHOTOGRAPHY 

beams as there are lenses placed behind the first one. The germ of this idea 
was first disclosed by Berthon and Audibert (Eng. P. 24,809/11). It was 
improved by Audibert in 1912 (Fr. P. 458,040; U.S.P. 1,124,253). In its 
final form (Eng. P. 355,835; U.S.P. 2,096,665) a divergent lens, with its nodal 
points well in front of the lens system, is placed in front of three normal lenses 
(Fig. 41). Here is the negative or divergent lens, whose nodal points are 
Ni and N2. The beam entering this lens is spread out slightly so that it will 




FIG. 41 

cover completely the field of the three lenses which lie immediately behind 0. 
These lenses are designated Li, L%, and L 3 . H. N. Cox has done considerable 
work in this field. In the first few of his patents, he placed a lens at one end 
of a cylinder (U.S.P. 1,645,374, and 1,645,417), and at the other end he placed 
a quadrant-shaped achromatic prism and filters. This prism projected four 
images in the space of a single frame of motion picture film. In a whole series 
of other patents Mr. Cox develops the idea proposed by the Audibert dis- 
closures. These patents are as follows: 



United States 


English 


1,699,226 


274,593 


1,700,252 


274,683 


i,735,io8 


310,533 


1,762,144 


314,546 


1,778,754 




1,781,496 




1,811,495 





Similar ideas were disclosed by J. A. Kienast (U.S.P. 1,843,642); M. de 
Francisco (U.S.P. 1,873,302); E. D. Goodwin (U.S.P. 1,921,918); L. M. 
Dieterich (U.S.P. 1,990,529); J. Szczepanik (Eng. P. 238,973); R. S. Alldridge 
(Eng. P. 322,801, 352,292, 352,293, and 352,294); and A. A. Arnulf and 
P. E. Bonneau (Eng. P. 418,562). A slightly different idea is disclosed by 
Paul Verola (Eng. P. 308,973). Immediately behind the lens (Fig. 42) there 
is placed a special trapezoidal prism of flint glass, with two sides made of 
crown glass. This divides the lens aperture into three parts, and the light 
through each part is refracted at a different angle, thus giving rise to three 
separated beams. 

All of these schemes appear to be very good solutions of the problem of easy 
and fool-proof color separation, but they make the calculation of lens correc- 



COLOR CAMERAS, TYPES 4 AND 5 



65 



tions an extremely difficult matter. For this reason, these systems suffer 
primarily from defects such as coma, etc. There seems to be always some 
group or other which utilizes them, but the group always appears to be in the 
experimental stage. At the present writing, the only commercially useful 
procedures for obtaining color separations appear to be restricted to a sys- 



T0W7L 




— Hint 



CrotVrt 



FIG. 42 

tern of light splitters behind the lens, a system which we have discussed 
in Chapter 5 under the heading of Type 3. Types 1 and 2 we can im- 
mediately discard as being theoretically unsound. Type 4 presents the 
problem of matching two or more lenses, besides the problems inherent in the 
invention of a light-splitting device which satisfies the two conditions im- 
posed by Do well. Therefore, these systems offer no advantages over those 
of Type 3. In Type 5, there is the problem of designing a color-corrected 
negative lens of wide aperture, which could be coupled with matched lenses 
to yield a plurality of images in the space usually occupied by a single image. 









CHAPTER 7 
THE PHOTOGRAPHIC EMULSION 



Th: 



.HE photographic negative material consists of a dispersion of micro 
crystals of silver iodobromide in gelatin. The term iodobromide is used to 
indicate that the composition of each grain is a homogeneous mixture of the 
two silver salts, the bromide and the iodide. Crystallographic studies have 
indicated that the structure within the individual grain is uniform, and that 
the distance of a negative halide from a positive silver ion is constant through- 
out each grain. But the grains are not identical to each other. They not 
only vary in size, but also in composition, and in their ability to react with 
light. It has been found that the faster grains were on the whole somewhat 
coarser and contained a greater percentage of iodide than the slower grains. 
It has also been found, that as the iodide content of a grain increased, the 
lattice distance between the ions increased. It is apparent, therefore, that 
the ordinary photographic emulsion is a complex mixture of light-sensitive 
elements, dispersed in a binding medium, such as gelatin. 

The purpose of the gelatin is primarily to insulate the individual grains from 
each other, thus allowing each grain to act as a unit by itself. The complete 
emulsion contains a large number of these grains, which exist in a statistical 
equilibrium. Uniformity in an emulsion is achieved only because the number 
of light-sensitive elements per unit volume is very great, and not because 
there is a uniformity of properties in the individual grains. This is very im- 
portant to understand, as several of the important characteristics of the photo- 
graphic material depend upon this point. It has been determined by Eggert 
and Biltz (Trans, Far. Soc, Vol. 34 (1938), p. 892) that approximately 5 per 
cent of the silver bromide grains in an emulsion of moderate speed are develop- 
able upon the absorption of but a single quantum of light energy. But the 
average grain in the same emulsion will require the absorption of approximately 
two hundred quanta of light before development becomes possible, which in- 
dicates a wide variation in the sensitivities. This emulsion contains almost 
three billion grains to the square centimeter of surface area. 

We can describe the negative material, then, as consisting of about three 
billion light-sensitive elements per square centimeter of emulsion surface, in- 
sulated from each other by a film of gelatin, and varying in their sensitivities 
to light over a wide range. When such an emulsion is exposed for a definite 
length of time, a certain number of the grains become developable. As the 
intensity of the light, or the time of exposure, is increased, a greater number 
of the grains becomes developable. The sole function of exposure to light is 

66 



THE PHOTOGRAPHIC EMULSION 67 

to make a definite number of the grains, per unit area of emulsion surface, 
developable. When a negative is exposed to the light that is reflected from 
an object, at each unit of surface of the negative there will be a certain number 
of grains made developable, the number depending only upon the intensity of 
the light which reaches that area. There is therefore a one-to-one correspond- 
ence between the intensity of the light which is reflected from any given point 
on an object, and the number of developable grains that is formed at the cor- 
responding point on the negative. 

When an emulsion grain is exposed to light, that grain is either developable, 
or it is not developable. This means that within the limits of fog-free de- 
velopment it is not possible to reduce to metallic silver a grain which has not 
received enough light to make it developable. What gives rise to density 
difference in a negative is the fact that a greater number of grams are being 
developed in one part of the emulsion than in another. It is true that as one 
develops longer, there will be a tremendous increase in the overall density, 
but this is not so much due to the fact that more grains are being developed, 
as to the fact that a greater proportion of each individual grain is being reduced. 

The grain itself is a mixed micro-crystal of silver iodide and silver bromide, 
with an average cross-sectional area of io" 8 square centimeters. This would 
make the width of an individual crystal not far from one /x (one thousandth 
of a millimeter). The average crystal will contain one billion, seven hundred 
million molecules of silver halides. The true function of development becomes 
quite clear from these figures. The average grain will become developable 
upon the absorption of two hundred quanta of light, and upon development, 
these grains can yield approximately one and a half billion atoms of silver. 
There is therefore an intensification factor of almost ten million, due to the 
technique of development. This is the real difference between the printing-out 
and developing-out emulsions. In printing-out emulsions, every atom of silver 
that is formed requires the absorption of one quantum of light. In developing- 
out emulsions there are ten million atoms of silver formed for every quantum 
of light that is absorbed. This explains also, the inherent slowness of dichro- 
mate printing, bleach-out, diazotype, and all other processes which use light- 
sensitive systems that yield the final image by the direct action of the light 
upon the sensitive elements. For this reason there may be a definite lower 
limit to grain size in fine-grain emulsions for negative purposes. The smaller 
the grain, the less the number of molecules of silver halides per grain, and 
therefore the less the number of silver atoms that may be formed by the 
complete reduction of such grains. It is silver that gives the density in a 
negative, so that if less silver is formed, less density results, and consequently 
a given amount of light yields a smaller quantity of final image. 

When an exposed negative is placed in a developing solution, the first thing 
that occurs is that the developer solution becomes absorbed by the gelatin. 
The rate at which this is accomplished depends upon the physical characteristics 
of the gelatin itself. This substance ranges in absorption.prbperties from being 



68 HISTORY OF COLOR PHOTOGRAPHY 

completely waterproof, to acting as a sponge. The gelatin used for photo- 
graphic, purposes falls somewhere between these two extremes, although it is 
much nearer the sponge stage than the other. The solution absorbed by the 
gelatin then reacts with the exposed grains, and here again there is a preliminary 
adsorption stage, which takes some time. There is therefore a time interval 
in which no actual chemical activity takes place. This is called the induction 
period, and it was used by Watkins as a measure of the time required for 
complete development to take place. The interval time is noted from the 
moment the negative is inserted in the developing solution, to the time that 
the image first makes its appearance. This interval, multiplied by a constant 
which is characteristic for each developer, denotes the proper time of develop- 
ment. This system of development is called the factorial system, and the con- 
stant is called the Watkins factor of the developer. In the days when pan- 
chromatic emulsions were not so common and negatives could be developed 
in red light, this system was a very handy one, but today, when the sensitivity 
of the emulsions is such that the development must proceed in total darkness, 
it loses its effectiveness. 

The chemical reaction by which the silver halides are actually reduced to 
metallic state, does not proceed as rapidly as the neutralization of an acid by 
a base. Rather the crystal is slowly eaten away. This means that a consider- 
able length of time must elapse before the grain is completely reduced. It is 
possible to stop the reaction at any point in this interval, and it is this char- 
acteristic that is used by technicians to vary the contrast of the finished nega- 
tive. When a material is developed to gamma infinity, it means that it is 
developed to a point where all the exposed grains are completely reduced. 

Consider a negative which has been exposed to a gray scale. In this gray 
scale let us suppose that the first step is completely transparent, therefore 
transmits ioo per cent of the light that is incident upon it. The second step 
we will consider as transmitting 50 per cent of the light, the third step 25 per 
cent, etc., each step transmitting half as efficiently as the step before it. In 
the exposed negative we will have a series of stripes which contain varying 
numbers of developable grains. It must not be supposed that the number of 
exposed grains in the first stripe will be just twice the number in the second. 
It must be recalled that the grains have a wide variation of sensitivities. The 
very first absorption of light will affect the extremely rapid grains, and the 
continued absorption will make the next fastest grains developable. The 
number of grains in each class is not the same, but is determined by the 
method used in the preparation of the emulsion. Fast negative emulsions 
contain a higher percentage of extremely fast grains. Such emulsions will 
give silver deposits upon the absorption of very small quantities of light. Slow 
emulsions, on the other hand, will not yield any visible amount of silver under 
the same conditions of exposure. Therefore, as the exposure intensity is de- 
creased, the number of grains containing a latent image may decrease at an 
entirely different tempo. 



THE PHOTOGRAPHIC EMULSION 69 

After the exposure, the negative is placed in the developer. In approxi- 
mately three or five minutes, there will appear an image in the first step, a 
very faint one. As the development time is increased, the image will appear 
in the other steps, while that in the first step will grow much stronger. We 
can analyze the process in the following manner. As soon as the induction 
period is over, the reduction of the exposed grains starts. As a first approxima- 
tion we can assume that all of the exposed grains begin to react with the de- 
veloper. Before the silver can become visible, it must be present in a certain 
minimum concentration. This can be achieved in two ways. There may 
be such a large number of grains made developable that when one-tenth of 
one per cent of each grain is reduced to metallic state, there will be sufficient 
silver precipitated to become visible. This is most probably the case in the 
first step of the gray scale. In the second and other steps, the concentration 
of the exposed grains per unit area of surface will be insufficient to yield a 
visible amount until a higher proportion of the individual grains is developed. 
Therefore no image will be visible in these steps until some time has elapsed, 
and the grain etched down to a much greater extent. In these cases, the dis- 
tance between grain centers is much greater than it is in the case of the first 
step, but the area of each grain of silver deposit is greater. Therefore the 
image in the first step would be considerably smoother than the images in 
the other steps, the difference corresponding to the differences that exist 
in pictures that are reproduced with a 40-line screen and those reproduced 
with a 60 or 120-line screen, in ordinary newspaper and magazine reproduc- 
tion. 

A silver deposit has a certain covering power; that is, it has a certain ability 
to absorb the light that falls upon it. If the grain 'is developed so that the 
area that the silver deposit covers is a relatively large proportion of the distance 
between grain centers, very little light will be transmitted by that image. If 
the development is so slight that only a very small portion of the space be- 
tween grains is covered, then the image will be very transparent. It is possible, 
therefore, to achieve a certain percentage transmission in two ways. The ex- 
posure can be made sufficiently large so that the distance between exposed 
grain centers will be very small. Then, upon but slight development, there 
will be obtained a coverage due to distribution of minute pinpoint deposits 
quite close to each other. Or the exposure may be very short, but the de- 
velopment quite long. In this case the same amount of silver will be produced, 
but the silver will be in the form of relatively large dots situated at relatively 
great distances from each other. 

This brings the discussion to the relationship which exists between exposure, 
development, and the resultant intensity of image. This intensity of image, 
really a measure of the concentration of silver that is deposited, can be meas- 
ured by its ability to absorb light. We define, first, several terms that are 
in common use. The ratio of the intensity of light that falls upon a silver 
image to the intensity of the light after it passes through that image, is termed 



70 HISTORY OF COLOR PHOTOGRAPHY 

the opacity. The logarithm, to the base 10, of this quantity, is termed density. 
Mathematically this can be written in the form 

Density = D = logarithm opacity = log O = log ±y- 

= - log y = - log f IOO y X l/lOO j 

= — log (% transmission) — log i/ioo 
= — log (% transmission) + 2.00 
= 2.00 — log (% transmission) 

/ is the intensity of the light transmitted by the silver deposit; I is the in- 
tensity of the light which falls upon the silver deposit. 

The term density is an extremely useful one. It denotes the degree of 
"blackness." Thus in a gray scale, as the steps increase toward the black, 
the density takes on higher values. A true black would have a density of in- 
finity, for such a black would have a transmission of zero per cent, and the 
logarithm of zero is minus infinity. A transmission of 50 per cent is equivalent 
to a density of 0.30, since the logarithm of 50 is 1.70. In the same manner 
it can be shown that a transmission of 25 per cent is equivalent to a density 
of 0.60, a transmission of 12^ per cent to a density of 0.90, etc. As the trans- 
mission is decreased by 50 per cent, the density increases by 0.30. In photog- 
raphy, the intensity of a silver deposit is denoted by its optical density. A 
high density in a negative corresponds to a heavy silver deposit at that point. 

It is possible to extend the use of this term to all surfaces which reflect or 
transmit light, in particular to the surfaces which are being photographed. It 
is not wrong, therefore, to speak of a given substance having a density equal 
to one, when we mean to say that the surface under consideration absorbs 
90 per cent of the light which falls upon it. This is especially true' of that 
small section of the picture which carries the gray scale. All pictures being 
photographed in color should carry a gray scale somewhere, for then the 
process of reproduction can be made into a science rather than an art. Let 
us center our attention upon this part of the picture. Upon exposure, there 
will be formed a series of areas in the negative in which there will be a pro- 
gressive increase in the number of developable grains per unit area of emulsion 
surface. The relationship between the different steps in the gray scale is 
known, and for the moment we can assume that step number one is com- 
pletely transparent or pure white, so that it transmits or reflects 100 per 
cent of the light that is incident upon it. Step number two will absorb 50 per 
cent of the light, step number three will absorb 50 per cent more than step 
number two, etc. Therefore, in the new nomenclature which we have just 
expounded, we may assign the density value of 0.00 to step number one, 0.30 
to step number two, etc. If the gray scale contains ten steps, these will be 
the density values: 



THE PHOTOGRAPHIC EMULSION 71 



Step No. 1 


0.00 


Step No. 


6 


1.50 


Step No. 2 


0.30 


Step No. 


7 


1.80 


Step No. 3 


0.60 


Step No. 


8 


2.10 


Step No. 4 


0.90 


Step No. 


9 


2.40 


Step No. s 


1.20 


Step No. 


10 


2.70 



The last step will transmit only 0.16 per cent of the light which is incident 
upon it, while the first step will transmit all of it. There will be a ratio of 
1 : 600 between the two, which is about ten times the ratio that one usually 
meets in practice. The ideal negative material would be such that there 
will be a continuous increase in the number of developable grains formed 
during the exposure in the entire scale of light intensities, ranging from 1.00 
to 600. Unfortunately this is not true, the usable range of most negative 
emulsions being but a fraction of this. 

Upon development of the negative, it will be found that it will contain a 
series of ten steps, which should correspond to the ten steps of the original. 
When the transmissions of these steps are measured and their densities deter- 
mined, it will be found possible to draw a curve in which the densities of the 
original scale are plotted horizontally, while the corresponding densities in 
the negative are plotted vertically, both in the same units. In a typical 
example, the density readings in a negative may be the following: 



Step No. 10 


o.35 


Step No. s 


1.40 


Step No. 9 


0.40 


Step No. 4 


1.70 


Step No. 8 


0.50 


Step No. 3 


1.85 


Step No. 7 


0.80 


Step No. 2 


1.90 


Step No. 6 


1. 10 


Step No. 1 


i-95 



The heaviest density in the negative is obviously due to that step in the 
original which transmitted the most light, that is step No. 1. Therefore, the 
density 1.95 in the negative will correspond to the density 0.00 in the original. 
It is customary to arrange the scale of densities so that they increase as we 
proceed upward along the axis which denotes the densities in the developed 
negative, and from right to left, on the axis denoting the densities in the 
original. Such a curve is depicted in Fig. 43. Density has been defined as 
being equal to a constant minus the logarithm of the percentage of trans- 
mission. But the percentage of transmission is identical to the intensity of 
exposure, therefore the density in the negative is directly proportional to the 
negative of the logarithm of the exposure. Now it is quite apparent why the 
densities in the original were plotted in the reverse direction to that which 
one usually plots in diagrams. 

An examination of the curve indicates that it can be divided into three parts. 
The first section is the part extending from the point E to the point A. Here 
the increase in the developed densities is not uniform for a corresponding in- 
crease in exposure. This section is called the toe of the curve, and it repre- 



72 



HISTORY OF COLOR PHOTOGRAPHY 



sents the region of underexposure. It can be interpreted to indicate that the 
number of developable grains per unit area does not increase regularly as the 
exposure intensity increases. It is only the extremely fast grains which are 
affected here, and the number of such grains is very small in proportion to 
the total number. Therefore, as the light intensity increases, only a slightly 
greater number of grains reach the developable stage, thus giving but a slight 
increase in the developed density. The increase in the number of developable 
grains becomes uniform by the time the light intensity reaches a value which 



<* "*> ^ 




2,70 2.40. Z.10 ISO ISO UQ .90 £0 .30 

Original Densities 

FIG. 43 



corresponds to the point A on the curve. Now a given increase in exposure 
intensity results in a definite and uniform increase in the density. This con- 
dition holds until we reach a light intensity that corresponds to the point B 
in the diagram. This section is called the straight-line portion of the curve, 
and the useful range of the negative material lies in the region encompassed 
between points A and B. 

The point A on the diagram corresponds to a density in the original of 2.00, 
while the point B corresponds to a density of 1.00. This means that the ex- 
posure intensity at B is just ten times that at A, The difference between A 
and B measured along the log-exposure axis is a measure of the latitude of the 
emulsion. In this case the emulsion was safe to use over a range of intensities 



THE PHOTOGRAPHIC EMULSION 



73 



of one to ten. This represents a very poor negative material since the light 
intensities in normal practice would have a ratio of one to fifteen or twenty 
for indoor photography, and one to forty or sixty for outdoor photography. 
In color work, the ratio would be' even greater when measured through the 
different filters. 

The third section of the curve is represented by the region between the 
points B and D. Here again there is no longer a uniform increase in developed 
density with a uniform increase in exposure. But this time, there is a slow 



-2.40 




2.7(7 ZAO 2. 10 f.80 t,50 \.*0 .90 

FIG. 44 



and continuous reduction in the rate of increase, whereas in the toe of the 
curve, there was a slow and continuous acceleration. It is not quite so easy 
to explain the reason for the existence of this upper limit. Undoubtedly, 
there is a considerable diminution in the number of grains still available for 
latent image formation when light exposures of this intensity are being used. 
Therefore, the number of grains that are affected no longer follows the simple 
rule expounded above. Then there is the possible effect of reversion due to 
the sharp increase in the products of the photo-chemical reaction. This also 
would reduce the rate of increase. These are possibilities that affect the number 
of developable grains that would be formed. 

There is one other possible explanation, which bears no relationship to the 
two mentioned above. It may very well be that even if the amount of silver 



74 HISTORY OF COLOR PHOTOGRAPHY 

deposit were in direct ratio to the exposure intensity, the opacity of the de- 
posit, hence the optical density, would still be below theory. In the normal 
negative emulsion, there are approximately thirty layers of grains present, 
and the grains in each layer are separated from each other by a thickness of 
gelatin. If every grain in the layer were completely reduced, there would be 
formed lamella of silver particles that are separated from each other by clear 
areas. The silver deposit will absorb a certain percentage of the light, it will 
transmit a definite percentage, and it will reflect the rest. It has been estimated 
that the covering power of a single layer of silver deposit was such that it 
would have a density of i.oo. This means that ten per cent of the incident 
light will be transmitted or reflected. Let us consider the light that is re- 
flected from the silver deposits. This will become incident upon other silver 
deposits lying either below or aboye the layer under consideration. If all 
thirty layers are completely reduced, there will still be some light reflected 
off the silver in the top layer which will be transmitted by the silver in the 
bottom layer, therefore there is a top optical density which can be measured 
from a lamellar silver deposit. This top density will depend, not on the amount 
of silver that has been reduced, but upon the ratio of the area which the silver 
deposit covers in each layer, to the total area of the emulsion surface. This 
ratio, as was indicated above, depends to a great extent upon the relative 
areas of the silver halide grains. ' The larger the grains, the lower will be the 
maximum obtainable density, for in that case there will be relatively large 
spaces existing between the individual silver specks which make up the image. 

Near this limit there could no longer be any effective increase in light ab- 
sorption, and the definite ratio which existed between the intensity of exposure 
and the optical density of the resultant image, no longer holds. In fact, this 
ratio begins to fall off, giving rise to the top or shoulder portion of the curve. It 
is very probable that the truth is in reality a combination of many causes, thus: 

i. At high exposure levels, there is a reduction in the number of grains which 
become developable, due to the fact that a very large proportion of the avail- 
able grains becomes used up before all the light is absorbed, and the remaining 
light affects grains of relatively low sensitivity. The concentration of these 
grains is less than the concentrations of the more sensitive grains. The failure 
of constant increase in this case is due to grain distribution. 

2. When light acts upon the silver halide grains, metallic silver is formed, 
and halogen gas is freed. The free bromine or iodine is supposedly absorbed 
by substances present in the emulsion or by the gelatin. But if the halogen 
gas be formed in large quantities, the gelatin or the other halogen acceptors 
may not be able to effectively remove it, and the photolytic reaction which 
forms the latent image may become retarded and finally brought to an equi- 
librium state. The failure of the constant increase is due in this case to the 
workings of the mass law. 

3. It is usually considered to be sufficient to make it developable if one de- 
velopment center exists upon a single grain. These development centers are 
formed around flaws in the grain surface. The methods of their formation 



THE PHOTOGRAPHIC EMULSION 75 

are still not well understood, but current opinion accepts the following explana- 
tion. The light is incident upon the entire surface of the crystal or grain. 
Wherever the light is absorbed, silver is formed photolytically. But this silver 
does not become anchored at the points where it is formed. It instead travels 
about until it becomes trapped by surface flaws on the grain. Here the silver 
accumulates to form a speck of rather large size (containing two hundred atoms 
for the average grain). This silver speck constitutes the development center. 
At extremely high exposure densities, many more than one sensitivity center 
may be formed upon each grain, some of these in the interior of the grain 
rather than on the surface. This will cause an effective reduction in the ratio 
of the light intensity and the number of developable grains formed. 

4. The concentration of developable grains at one spot may be so great 
that during development there is a tremendous increase in halide content, due 
to released halide ions formed by the developer reaction. This will increase 
the halide ion concentration locally to a point where it will stop the chemical 
reaction which is taking place. This again is an application of the mass law, 
but now the application is to the secondary chemical reaction rather than to 
the primary photolytic reaction. The two cases must be carefully distin- 
guished. In the first case it is the intensity of the latent image that is reduced, 
while in the second case it is the development reaction which is reduced. 

5. There is no diminution in the intensity of the latent image formation, 
nor is there a reduction in the intensity of the silver deposit formation. In 
fact, there may be no diminution in the rate at which silver is formed. What 
is lowered is the covering power of the deposit, since that covering power de- 
pends (at this concentration of deposit) not so much upon the concentration 
of the silver present at any point, as upon the extent to which it covers the 
area in each layer. 

The curve in Fig. 43 has other characteristics than the ones discussed above. 
The straight-line portion of the curve starts at the point A and ends at the 
point B. In this region there is encompassed a brightness ratio of 1 to 10, 
which is equivalent to a density differential of 1.00. This is a differential in 
the exposing light. When the negative was developed, we found that this 
density differential in the exposure light gave rise to a density differential of 
1. 00 in the silver deposit. This ratio, the relationship between the logarithm 
of the exposure and the density in the resultant silver image, is termed con- 
trast. In our example, the contrast has a value of 1.00, since a density dif- 
ferential of 1. 00 in the original gave rise to a density differential of 1.00 in 
the duplicate. But if we had developed the negative but three-quarters of 
the length of time used to obtain the negative depicted in Fig. 43, we would 
have obtained a set of densities such as these: 

Step No. 10 0.20 Step No. 5 1.10 

No. 9 0.30 No. 4 1.30 

No. 8 0.50 No. 3 1.50 

No. 7 0.70 No. 2 1.53 

No. 6 0.90 No. 1 1.55 



76 HISTORY OF COLOR PHOTOGRAPHY 

The curve in Fig. 44 represents this result. Now we notice that the straight- 
line section of the curve starts at a density of 0.70 on the log exposure scale, 
and ends at a density of 2.40, so that the range of brightness which is correctly 
portrayed by the negative bears the ratio of 1 to 50. Corresponding to this 
density differential of 1.70 along the exposure axis, the negative deposited a 
density differential of 1.20. The ratio between the two is 1.20/ 1.70, or 0.70. 
This is the contrast of the negative development, in the second example. From 
this point of view, it is seen that contrast is the degree to which a negative 
duplicates the brightness range of the original. This can be carried over into 
the positive process to indicate the degree to which a positive duplicates the 
brightness scale of the negative. Since we are dealing with straight lines, 
this ratio becomes identical to the tangent of the angle which the straight- 
line portion of the curve makes with the exposure axis. This value of the 
tangent is normally designated by means of the Greek letter y (gamma), 
therefore this letter has been universally adopted as a designation of contrast. 
Development is usually denoted as taking place to a gamma of, say, 0.80. By 
this is meant, that the development has been carried out until the brightness 
scale in the negative is 80 per cent that of the original. Up to very recently 
it has been customary to develop negatives to a value of gamma equal to 0.80, 
but with the advent of the more contrasty fine-grain emulsions, it is becoming 
increasingly more popular to push the value more closely to 1 .00. 

We have stated before that the idea of contrast can be applied to the posi- 
tive as well as to the negative. Consider, then, a positive that has been de- 
veloped to a contrast of 1.50. This means that the positive reproduces the 
brightness scale in the negative to an exaggerated degree; it intensifies it by 
a factor of 1.50. If the negative had been developed to a gamma of 0.80, then 
what would be the relationship between the brightness- scale in the original 
and in the final reproduction? Obviously it will be the product of the two 
gammas, that is 0.80 X 1.50, or 1.20. The reproduction will intensify the 
brightness scale of the original by a factor of 1.20. This is a general rule. 
When a reproduction is made by a series of steps, each step being capable of 
its own contrast scale, then the overall contrast will be equal to the product 
of the individual contrasts. In color reproduction, there will be at least three 
steps, each of which is capable of changing the brightness scale. The first 
step will consist in the making of negative separations. This will usually be 
carried out to a contrast of 0.80, so that a density differential of 1.70 (the 
usual amount in outdoor photography), will be condensed to 80 per cent of 
this value, or to a differential of 1.36. The negatives will next be converted 
into positives, and here again the contrast may be varied at will. The final 
step is the conversion of the silver in the positive to color. This will again 
have its own conversion factor, and a density of 1.00 in silver may give a 
color density that is either a fraction of, or many times this value. In the 
iodide mordant processes, the silver density is multiplied by at least four when 
converted into a mordant and dyed with a basic dye. In metallic toning 



THE PHOTOGRAPHIC EMULSION 77 

processes, the silver density is usually converted into a color density that is 
not far from the original value. In the first case, if it is desired to maintain 
an overall contrast of 1.00, the silver density must be developed to a gamma 
that is one-fourth the gamma required in the second case. 

Another very important characteristic of the photographic emulsion, is the 
speed. This is denned as the least amount of light that must be used before 
a suitable image is obtained in the negative. It is only the straight-line por- 
tion of the curve which is of interest to the photographer, since he must re- 
produce a whole series of light intensities, rather than a single one. Therefore 
any measure of the emulsion speed must take this into consideration. There 
are several methods for the designation of film or emulsion speeds, and of 
these, only one method is of value. This is the method based upon the re- 
searches of Hurter and Driffield. They designated the speed by the value of 
the intercept of the straight-line portion of the curve with the exposure axis. 
The Weston system of speed rating is directly proportional to this value. There 
is another method which is very popular in Europe. This method is of value 
to the astronomer, and was in fact developed by Scheiner, an astronomer. 
In this scheme, the emulsion whose speed is to be determined is exposed be- 
hind a gray scale, whose steps differ from each other by a value of 0.10 in 
density. The steps are numbered 1, 2, 3, etc., so that the most transparent 
step bears the designation 1, and the other steps with a smaller transmission, 
bear the larger numbers. The speed rating of an emulsion is the number 
which is just visible after it has been exposed and developed in a definite 
manner. An emulsion with a rating of 18 is exactly half as fast as one with 
a rating of 21, and this is half as fast as one with a rating of 24. 

It is easy to see that speed rating according to this scheme is completely 
false as far as photography is concerned. Consider the curves for two emul- 
sions, both developed to the same gamma, and both having their straight- 
line portions intercept the exposure axis at the same point. Photographically 
the two would be identical, and the reproduction obtained by the use of either 
of these emulsions, will be the same. But suppose that one of the emulsions 
has a toe which differs considerably from the toe of the other. One toe may 
be extremely short, and have a slope that is only slightly different from the 
main portion of the curve. The other may have a toe which is very long, 
and which has a very much lower inclination to the axis than the first. In 
the Scheiner system, this second emulsion would be given a speed rating which 
may be three or six units higher than the first. There will be a tendency to 
considerable underexposure if the Scheiner rating is used. In Germany, the 
emulsion speed is determined by a modified form of the Scheiner method. 
The speed is determined from that point on the curve which gives a density 
which is 0.10 above fog. The exposure and development are carried out under 
stated conditions. There is very little difference between this, the DIN speci- 
fication, and the European Scheiner. The American Scheiner ratings are deter- 
mined in a manner that is completely analogous to Weston or H & D speeds. 



78 HISTORY OF COLOR PHOTOGRAPHY 

For this reason, great care must be used not to confuse American Schemer 
with the European. 

In the early days of photography it had been assumed, and most probably 
verified by the crude experimental procedures that were available at that 
time, that a reciprocal relationship existed between the intensity of exposure, 
the time of exposure, and the density of the final image. This relationship 
could be expressed by the equation 

I X T = constant. 

/ is the intensity of exposure and T is the time of exposure. This may be 
explained to mean that if the intensity of the exposing light is decreased by 
50 per cent, it will be necessary to increase the time of exposure by a like 
amount, in order to obtain the same result in image intensity. An exposure 
of five seconds at/: n is identical to an exposure of twenty seconds at/: 22, 
since the intensity of exposure is four times that which it is at/: 22. This is 
known as the Reciprocity Law. The general validity of this law was first 
questioned by Abney, and experimentally found to be not true by Schwarz- 
schild. No adequate explanation has yet been made to explain why reduction 
in exposure intensity followed by an exact mathematically equivalent increase 
in exposure time, should yield an image density that is sometimes considerably 
lower than the one made at a higher intensity level. But the truth of this 
anomaly has been well established. 

Consider two emulsions, the first of which has a rating of Weston 40, while 
the second has a rating of Weston 4. We will suppose that we are making 
color separations from a still object, such as an oil painting. The object is 
illuminated in the normal manner, by means of two lights situated so that the 
lines from their centers to the center of the object, make 45-degree angles 
with the plane of the object. The intensity of illumination is such that it will 
require an exposure of approximately twenty seconds through the red filter, 
when using the emulsion with a rating of Weston 40. After exposure and de- 
velopment, it is found that a very flat negative results, indicating that the 
fast emulsion is not the proper material to use. The slow material will have a 
much higher contrast, and the apparent flatness can be overcome by a higher 
development gamma. In accordance with the relative speed ratings of the 
two emulsions, it would be supposed that an exposure ten times that of the 
other would be required. If that technique were followed, it would be found 
that the negative would be considerably overexposed. It will require an ex- 
posure only three or four times that given the first material to yield a usable 
negative. 

This can be answered by the statement that the reciprocity law failed. 
The rating Weston 40 has been assigned to the first emulsion by a speed deter- 
mination in which the intensity of illumination was such that an exposure of 
a fraction of a second was required. This will be identical to the conditions 
under which the emulsion would be used in actual practice. The speed rating 



THE PHOTOGRAPHIC EMULSION 79 

of Weston 4 was assigned to the slower emulsion by a test in which the in- 
tensity of illumination was quite low in comparison with the first, and ex- 
posures in the neighborhood of a minute or so, were used. These are the 
conditions for which this type of emulsion is intended. But when the first 
emulsion is used under the conditions intended for the second, the speed rating 
can no longer be applicable. A new determination is required, in which the 
lower intensity of illumination should be used. Were a test made under these 
conditions, it would be found that the speed rating would have been more 
nearly Weston 14 than 40. 

A more practical effect of the failure of the reciprocity law may be had by 
a test of the extremely fast modern emulsions, under different lighting condi- 
tions. With these emulsions, the failure of the law is quite evident even with 
a drop in intensity levels of four to five times. The real speed of these emul- 
sions is made apparent only when the exposure time is of the order of one- 
fiftieth of a second or less. When the exposure becomes as high as a tenth 
of a second, a loss of speed will become quite evident. It is this failure, which 
has given rise to the skepticism of the professional photographer to the speed 
values of the American film manufacturers. The distrust of the European 
speed ratings is due to a more fundamental fault, which was discussed above. 
It is their insistence upon using threshold rather than actual values for the 
speed. The photographer has no interest whatsoever in the ability of the 
emulsion to yield a density of 0.10 above fog, if that density lies in the region 
of underexposure or the toe of the curve. Up to the present writing, no emul- 
sion has been made where the straight-line portion of the curve begins at a 
density value less than between 0.30 and 0.40. From this discussion, it be- 
comes evident that an emulsion must be used only for the purposes for which 
it is intended, and at the light intensity level for which it is intended. 

The last characteristic of the emulsion material in which the photographer 
is interested, is its spectral sensitivity. The silver iodo-bromide grain is light- 
sensitive only to the ultraviolet and blue regions of the spectrum. The emul- 
sions that contain the greatest amount of iodide will have their sensitivities 
extend slightly into the blue-green region, say up to 540 m/i. But the sensi- 
tivity of the emulsion beyond the 500 mark is very much lower than it is up 
to there. The peak of sensitivity of such emulsions, when exposed to light of 
sunlight quality, is at 480. At 520 the sensitivity is about one-tenth of what 
it is at the peak, and falling very rapidly. 

It has been found possible to add dyes to the emulsions just before they are 
coated upon the glass or film base, with the result that the sensitivity of the 
resultant material becomes extended to cover the absorption range of the 
dyed silver halide grain. This almost, but not quite, coincides with the ab- 
sorption range of the pure dye. For this reason it has been inferred that a 
new light-sensitive material is formed by the action of the dye upon the silver 
halides, and it is this light-sensitive material that gives rise to the latent image. 
The number of dyes available for this purpose has been extended tremendously 



80 HISTORY OF COLOR PHOTOGRAPHY 

and it is possible to make emulsions whose sensitivity lies in almost any de- 
sired range. While it is true that with the extension of the sensitivity to re- 
gions beyond that of the absorption of pure silver iodo-bromide the sensitivity 
to the ultraviolet and the blue becomes diminished, it has not been found 
possible to completely reduce this native sensitivity to any considerable degree- 
In all cases of induced sensitivity, the speed of the emulsion to the blue rays 
remains considerably greater than the speed to the rays in the induced range. 
But this is neutralized to a considerable extent by the fact that even daylight, 
which is the bluest white light at our disposal, contains a tremendous excess 
of red and green over blue, and when an emulsion is exposed to such light, 
the intensity of red and green light exposure is several times that of the blue. 
Experiments upon the absolute sensitivity of panchromatic emulsions to light 
of selected wavelengths have shown that from 60 to 90 per cent of the total 
sensitivity of these emulsions still lies in the blue and ultraviolet regions, while 
the rest of the sensitivity is divided among the remaining two primary color 
regions. 

In so far as color sensitivity is concerned, we can divide photographic emul- 
sions into four classes. The first will contain those emulsions whose sensi- 
tivities coincide with the native sensitivity of the iodo-bromide grain. These 
emulsions are called blind, or non-colorsensitive, and their activity is limited 
to the ultraviolet and the blue end of the spectrum. The second class con- 
sists of the orthochromatic emulsions, those which have been sensitized to 
the green and the spectral yellow region. These emulsions extend their ac- 
tivity up to 600 mju. Some of the older types of orthochromatic emulsions 
have their limits at about 560 m/x, but the later tendency is to sensitize for 
the entire green primary range, up to 600. The third class of emulsions is 
termed panchromatic, and as the term implies, these emulsions are sensitive 
to practically the entire visible range, up to 650 m/x. The visibility of the 
normal eye is very low beyond this point, so it is sufficient for most purposes 
to photograph only up to 650. The fourth, and last, class contains the emul- 
sions that are sensitive to the infrared rays, beyond 700 m/x. These have but 
little application in color photography. 

The ability of the emulsions to be sensitized to color depends to a great 
extent upon their composition and upon the size of the individual grains. 
In the main, the smaller grains sensitize much more readily. Also the grains 
with a minimum of iodide sensitize most readily. Unfortunately these are 
the conditions for slower emulsions. It is not possible then, to prepare an 
emulsion of the greatest possible sensitivity, and increase that sensitivity by 
the addition of dyes. To make color-sensitive emulsions that are very fast, 
it becomes desirable to make the individual grains quite fine, and with a 
minimum of iodide. It is perhaps for this reason that the blue sensitivity of 
extremely fast emulsions is appreciably lower than normal. It has also be- 
come the practice to prepare even the blue-sensitive emulsions in this manner; 
that is, to prepare the basic emulsion so that it consists of fine grains with a 



THE PHOTOGRAPHIC EMULSION 81 

low iodide content, and to increase this sensitivity by the addition of yellow 
dyes which sensitize for the blue. This serves a dual purpose. First of all, 
it restricts the sensitivity to the blue and the ultraviolet, so that such emul- 
sions can be used for the blue densities separation without filters. Secondly, 
the amount of blue light that is transmitted by the emulsion becomes reduced 
to an appreciable extent, so that the use of such an emulsion for the front 
element of a bipack, will make unnecessary the use of a yellow filter to remove 
the unused blue light before it reaches the rear element. We will discuss 
this to a greater extent in the next chapter. 



CHAPTER 8 
TRIPACKS AND BIPACKS 



Wh 



HEN color separations are made by means of a one-shot camera uti- 
lizing any one of the light-splitting devices which were discussed in the pre- 
vious chapters, there is occasioned the loss of the greatest part of the light 
intensity. First of all there is a loss of at least 25 per cent of the light caused 
by absorption by the mirrors which make up the light splitter. Then there 
is a further loss of at least 66 per cent of the remaining light due to the fact 
that the lens beam is divided into three equal parts, and each sub-beam is 
used to make an exposure through a filter. Therefore only one quarter of the 
original intensity is present in each beam. But this does not constitute all 
the losses. The standard camera is designed to use the A, B, and C5 Wratten 
filters. Of these, the A filter is the most efficient, transmitting almost 80 per 
cent of the available red light. The other two are very inefficient, since they 
transmit less than 40 per cent of the other primaries. It is not difficult to 
calculate the efficiency of a one-shot camera which uses these filters. Let us 
suppose that we will use the fastest negative material that is available, and 
that this material has the following factors for the three filters when exposed 
to Mazda light: 

A filter 4,0 

B filter 6.0 

C filter 1 0.0 

In order that a balanced set of separations be made, it 'is necessary that the 
material behind the B filter be given one and one-half times the exposure given 
the material behind the A filter, and the C filter separation requires two and 
one-half times this exposure. But in a one-shot camera it is not possible to 
give to one negative a greater exposure time than to another, so that the 
difference in exposures must be made by dividing the light in such a manner 
that the beam which is deflected to the B filter has one and one-half times, 
and the beam which goes to the C filter has two and one-half times, the in- 
tensity of the beam which is directed to the A filter. To achieve a correctly 
exposed set of negatives, it is necessary to direct 50 per cent of the light to 
the blue filter, 30 per cent to the green, and 20 per cent to the red filter. 

It is possible to get some idea as to the amount of light that is lost in such 
a system by a determination of the number of times by which the normal 
exposure must be multiplied, in order to get a correct exposure by this means. 
The image behind the red filter receives but one-fifth of the total available 

82 



TRIPACKS AND BIPACKS 83 

light which enters the lens. This introduces a factor of five. Of this light, 
most is lost by the passage through the filter, which has a factor of four. 
There is therefore a lens factor of 20. A similar reasoning will demonstrate 
that this factor is true for the images behind the other two filters. A one- 
shot camera, working at maximum efficiency and utilizing the standard A, 
B, and C5 filters, will have a lens factor of 20. If the negative material that 
is used has a Weston speed rating of 96 for Mazda light, then the camera will 
have a speed rating of almost 5. This may be somewhat improved by using 
a different set of filters and special materials behind them, but we will leave 
a discussion of this for a later chapter. 

The great inefficiency of a light splitter has given rise to efforts which will 
make their use unnecessary. One device which accomplishes this result is 
the tripack. This consists of a combination of three films, one situated on 
top of the other, and each coated upon its own base. Two of the films are 
placed with their emulsion sides in contact with each other, in the form of a 
bipack. The third emulsion is placed behind this, with the emulsion side in 
contact with the rear of the bipack. There is a separation between them equal 
to the thickness of the base of the center emulsion. This can be made as 
thin as 0.003 inch, but even this small amount is sufficient to give a slightly 
larger image, when the enlargement is several diameters. 

The three emulsions comprising such a pack, must fulfill certain definite re- 
quirements as far as color sensitivity is concerned. The front emulsion must 
be color-blind, that is, it must be sensitive only to the blue rays of the spectrum. 
The native sensitivity of an emulsion that has received no treatment with 
dyes has its maximum at 480 mju, after which it declines rather sharply so 
that at 520 mix the sensitivity is only 10 per cent of what it is at the maximum. 
At 540 the sensitivity has declined to one per cent, so that for all practical 
purposes it can be assumed that the sensitivity of the blind emulsion has a 
limit at 530. The C5 filter transmits most efficiently at 440 mjtf, and at 520 
it has a transmission that is approximately 5 per cent of the maximum. There 
is not a tremendously great difference between the two, although for really 
fine work the excess exposure in the green range occasioned when a color- 
blind emulsion is used without any filter, would result in a falsification of the 
blue-greens. They would be rendered slightly greener or yellower. This film 
is the front element of the bipack, so that its base faces the lens, and its emul- 
sion side is in contact with the emulsion side of an orthochromatic material. 
It is possible to choose a material such that it will have an almost uniform 
sensitivity from 500 to 580 m/z, after which there will be a very sharp decline, 
so that at 600 m/z the sensitivity will be practically negligible. It is customary 
to stain the emulsion side of the front element with a yellow dye, in such a 
manner that the yellow stain does not penetrate deeply into the depth of that 
emulsion but still is present in sufficient intensity to absorb any blue light 
which may reach the top emulsion surface of the front element. This will 
prevent the blue rays, that have not reacted with the silver halide grains to 



84 HISTORY OF COLOR PHOTOGRAPHY 

form a latent image, from reaching the middle and rear emulsions. Thus no 
blue light can act here to destroy the quality of these separations. Since the 
orthochromatic emulsions can be made to have very sharp cuts, the green 
separation made by this means should be every bit as good as a separation 
made behind a B filter. 

The base of the orthochromatic emulsion is customarily stained red to ab- 
sorb any green rays which have not reacted to form a latent image in the 
central film. This prevents these rays from registering in the rear element, 
and' thus spoiling the quality of the red filter separation. This rear element is 
a panchromatic emulsion but since only red rays can reach it, only the red 
record will be registered upon it. Here again, it is seen that the quality of 
the separation must be every bit as good as the one obtained behind a red 
filter. The tripack system suffers only in the quality of the blue filter separa- 
tion, and even here, the loss is not very great. 

One defect of this system, as has already been pointed out, is that the rear 
emulsion lies several thousandths of an inch behind the plane of the other 
two emulsions. This gives rise to a slightly larger cyan image, which would 
be noticeable only at enlargements greater than three diameters. Where no 
such enlargement is desired, the differences in image size are negligible. But 
with this increase in image size there is also a lack of definition, and this is 
of greater consequence. The diffuseness of the rear image is highly aggravated 
by the fact that the emulsions are not optically clear materials which transmit 
completely the light which they do not absorb. The individual silver halide 
grain is a substance which has an index of refraction, relative to gelatin, 
whose value is 1.50. This means that approximately 10 per cent of the light 
will be reflected from the grain-gelatin interface. This reflection causes a cer- 
tain amount of image diffusion to take place, for at the higher intensities the 
reflected light will form latent images upon grains which do not lie within 
the geometry of the image being photographed. Another cause of poor defini- 
tion lies in the fact that the grain does not absorb all the light that passes 
through the interface. A considerable portion of the light is transmitted to 
the grains below. Here again the high index of refraction plays an important 
role, since it causes the transmitted rays to deviate from a straight line. But 
the diffusion introduced by this is considerably less than that produced by 
the scatter from the grain-gelatin interfaces. 

The ideal type of tripack would consist of three emulsions arranged, as in- 
dicated above, in the form of a bipack plus a single film behind the bipack. 
The emulsions that constitute the bipack should be completely transparent 
to the colors to which they are insensitive, and completely opaque to the rays 
with which they react. In the normal form, therefore, the first emulsion would 
be sensitive to the blue rays only. It would be completely transparent to 
the red and the green primaries, but completely opaque to the blue. There- 
fore no blue rays could penetrate beyond this emulsion. The second, or middle, 
emulsion would be sensitive to the green, in which case it would be completely 



TRIPACKS AND BIPACKS 85 

transparent to the red rays, and completely opaque to the green. The rear 
emulsion would be sensitive to the red. This emulsion need not be trans- 
parent, nor need its sensitivity be limited to the red, for since the blue and 
the green rays are completely removed before they reach this emulsion, they 
cannot register any image here. A pack such as this would be extremely fast 
and efficient in the use of the light. The blue component of the lens beam 
would register its image in the first layer, the green component would register 
in the second layer, and the red component in the last layer. No light would 
be lost, since each layer is completely transparent to the light to which it is 
not reactive. 

Let us compare this with practice. The normal negative emulsion consists 
of particles of silver iodo-bromide that have been color sensitized by treat- 
ment with certain types of basic dyes. The grains have a very high index 
of refraction relative to gelatin, so that when light is incident upon the top 
layer of grains in the emulsion, approximately 10 per cent of the light is re- 
flected off. This 10 per cent cannot form an image upon the grain which lies 
within the geometrical limits of the image, but can form an image upon grains 
that lie outside of this. Upon its passage through the entire depth of an emul- 
sion, there will be approximately thirty layers of silver halide grains upon which 
the transmitted light will be incident. The net result is that a considerable 
portion of the light will be lost by scatter. Some preliminary tests indicate 
that an unexposed emulsion has a density of 0.60 to white light, and a value 
somewhat lower than this, about 0.40, to light to which, theoretically at least, 
it should be transparent. This means that an ordinary negative emulsion will 
lose by scatter from the grains and absorption by the gelatin at least 60 per 
cent of the light that falls upon it. This loss of light effectively reduces the 
advantages of the tripack. Consider a practical case. We will use the fastest 
possible material for the rear element, a material to which we can ascribe a 
red filter factor of 2. This means that when exposed to Mazda light the 
emulsion will have 50 per cent of its entire sensitivity lying in the red region. 
Up to 1940, the fastest emulsion had a Weston speed rating to Mazda, of 
approximately 100. The red sensitivity can be assigned a value of 50, since 
the red filter factor is 2. But this emulsion is to be used behind two other 
emulsions, and each of these will transmit only 40 per cent of the light which 
is incident upon it. Therefore only 16 per cent of the original light intensity 
will reach the rear element, and the speed rating will be only 16 per cent of 50, 
or 8. The best possible tripack can have a speed rating which is only 60 per 
cent higher than the best possible one-shot camera which uses the standard A, 
B, and C5 filters. This slight increase in speed does not in itself pay for the 
diffusion and lack of definition that is inherent in tripacks. However, tripacks 
can be used in ordinary cameras, necessitating only a different type of plate- 
holder, whereas one-shot cameras are intricate, cumbersome, and very delicate 
instruments. For these reasons, the interest in tripacks has never waned, and 
efforts are continuously being made to improve them, and to make the front 



86 HISTORY OF COLOR PHOTOGRAPHY 

emulsions more transparent so that better definition and more speed can be 
obtained. In the United States, there is at this writing one tripack being mar- 
keted, and this has a rating of Weston 3. 

The first disclosure concerning the possibility of this system was made by 
Ducos du Hauron, who was responsible for so many innovations in color 
photography. Du Hauron ("Handbook to Photography in Color," p. 10) 
fully appreciated the difficulties that such a system presented. In his book 
and in his patent disclosures (Fr. P. 250862) he discussed these problems. 
He pointed out the necessity for maintaining good contact between the three 
elements so that no anistropic media be interposed between them. The exist- 
ence of such media would induce a loss of definition. He described the neces- 
sity for the use of transparent or almost transparent front elements, so that 
scatter be reduced to a minimum. Scatter is the cause of the greater part of 
the image diffuseness. He was also well aware of the fact that the rear emul- 
sion lay a finite distance behind the image plane of the lens system, and 
that this gave rise to a slightly larger image as well as a slightly diffused one. 

To offset the possible effects of poor contact between the various members 
of the tripack, J. H. Smith coated the emulsions directly one on top of the 
other, but with an insulating layer of collodion between them. In this manner 
there was obtained a very compact pack, since the collodion layers could be 
made as thin as desired. Filter dyes could be placed in this insulating layer, 
so that there was removed the possible interaction between filter dye and 
sensitizer. The presence of the intermediate dyed collodion layers made it 
possible to strip the component emulsions apart, and process them individually. 
This scheme is disclosed in United States patents 781469 and 886883; German 
patents 165544 and 185888; and English patent 19940/04. The idea crops 
up again in a disclosure by S. Schapavoloff (Eng. P. 205807) and more 
recently by W. T. Tarbin (U.S.P. 1871479). 

Biehler (U.S.P. 2088145) accomplishes the same result by coating inter- 
mediate layers of soluble cellulose esters between the emulsion layers. The 
surfaces adhere to each other merely by surface adhesion forces. The Gevaert 
Company (Eng. P. 481702) would separate the emulsion layers by means of 
paraffin. The melting point of the paraffin film separating the two upper 
emulsions would be quite a bit different from that of the paraffin layer sepa- 
rating the central from the bottom emulsion. In this manner it is an easy 
matter to strip the emulsions after exposure and process them individually 
after they have been transferred to other supports. Rowland S. Potter (U.S.P. 
2126137) placed a thin transparent adhesive consisting of a soap solution, 
between the layers. This permitted the pack to be processed as a single unit, 
since the soap film was permeable to aqueous solutions. The members of the 
pack could be separated after the processing was finished. 

E. A. Lage (Eng. P. 183189) applied strong pressure to the pack. L. von 
Tolnay and L. von Kovodsznay applied a vacuum to bring the members into 
close contact. E. A. Weaver (U.S.P. 1951043) would put a mild adhesive 



TRIPACKS AND BIPACKS 87 

between the elements, the adhesive containing at least 60 per cent glycerin, 
so that as far as the ray was concerned, the complete unit, gelatin-adhesive- 
gelatin, would constitute an isotropic medium. Since the glycerin content 
of the paste was so high, there would be no tendency for any of it to penetrate 
into the depth of the emulsion. W. H. Moyse (U.S.P. 1900459) claimed to 
have obtained better contact by the mere expedient of surface coating the 
emulsion layers with a colloid. The one which was placed upon the front 
element could contain a yellow dye. 

But the problem of good optical contact can be solved most easily by the 
use of pressure plates. This is the procedure adopted by the Defender Com- 
pany, who market a commercial tripack. The three elements are placed be- 
tween two pieces of glass of appropriate thickness, which is determined by 
mechanical requirements. The sandwich is then loaded into the plateholder, 
the back of which has heavy springs which exert considerable uniform pres- 
sure upon the glass plates. This simplest of all devices proved eminently 
successful and altogether sufficient to overcome any flaws that might be due 
to poor contact. 

A much more serious problem was the lack of balance that existed between 
the three elements unless special precautions were taken to prevent it. This 
lack of balance arose from two possible sources. The rear emulsions, having 
substances in front of them that lose a considerable portion of the light, are 
less fully exposed than the front. Shadow detail is much less defined, there- 
fore, since these sections of the image would invariably fall upon the toe or 
underexposure side of the characteristic curve. As was indicated above, only 
16 per cent of the light which enters the lens reaches the rear element. To 
obtain a perfect balance, it is required therefore to use an emulsion for the 
rear whose red sensitivity is at least six times as great as is the blue sensitivity 
of the front emulsion, and at least three times as great as the green sensitivity 
of the intermediate film. It is not a difficult matter to find emulsions whose 
sensitivities are at the required levels. The process emulsions have sensi- 
tivities that are one-tenth that of the slower negative emulsions, and it is 
possible to obtain any desired intermediate value. A certain prominent film 
manufacturer makes a fast negative material with a relative rating of 250. 
He also makes a moderate speed material rated at 150, a commercial type 
with a rating of 100, and at least three others whose values range from 10 to 
80. But there is one bad, feature of utilizing a mixture such as this. As the 
speed of the emulsion goes down, its contrast goes up, and its latitude be- 
comes quite narrow. This is the second cause of poor balance. If we have a 
pack whose front emulsion has a relative blue speed of 10, a central film whose 
green sensitivity has a value of 25, and a rear material with a rating of 60 for 
the red, we would have a pack balanced for speed. But when the elements 
are developed it will be found that the front negative will be considerably 
more contrasty, and the rear negative considerably flatter than the central 
one. It is also a question whether by changing the development, both as to 



88 HISTORY OF COLOR PHOTOGRAPHY 

time and composition, it would be possible to really balance the three for 
contrast. And unless this balance is maintained, falsification of color repro- 
duction will be obtained. 

This phase of the problem was recognized by F. E. Ives (U.S.P. 1 173429). 
To maintain a correct contrast balance he insisted that it was essential that 
the same basic material be used for all three elements. The blue-sensitive 
emulsion would be non-sensitized, the other two sensitized for their region 
only. E. J. Wall ("History of Three-Color Photography," p. 163) points out 
that this device falls short of obtaining the correct result, since the contrast 
scale varies with the color of the light that is used for exposure. The silver 
halide grain itself has a greater absorption of the blue rays than of the other 
two primaries, so that the image due to the blue rays is situated on top. This, 
to a considerable extent, reduces the contrast. The red and green also give 
different gammas when developed for the same length of time, but to a very 
much more moderate extent. But this argument loses force because the dif- 
ferentials in the contrasts that are due to this cause are well within the range 
of the material. It is possible to equalize the blue contrast by giving the blue 
filter negative from 10 to 100 per cent more developing time. The green 
negative requires at most 10 per cent additional development. It is important 
to note that Ives brought the three contrast scales to a point where their 
balance became possible. 

While this solved the problem of contrast balance, it left completely open 
the problem of speed balance. F. Thieme (Ger. P. 163282) reduced the speed 
of the front emulsion by staining it with acridin orange N. A. von Biehler 
(U.S.P. 1994627) adopted a similar idea. A highly sensitive basic emulsion 
was used for all three elements. The base of the film carried the emulsion on 
one side and a thin gelatin layer on the other. This layer contained one 
gram of tartrazine per 100 grams of gelatin if it was coated upon the blue- 
sensitive element. Thus the blue sensitivity was maintained at the same high 
level, but the intensity of the blue rays in the exposing light was reduced to 
a point where its action would equal the action of the other colors. The middle 
emulsion was sensitized to the green by means of erythrosin or pinafiavol. 
The base of this emulsion carried a gelatin layer with 8 grams of tartrazine 
per 100 grams of gelatin. The rear element was sensitized to the red by means 
of pinacyanol blue. 

In order for the heavy yellow backing of the central element to screen the 
green sensitive emulsion from the blue rays, it is necessary for this emulsion 
to be loaded with its carrier side to the lens. The bipack is now the rear 
element of the tripack, rather than the front. This means that the central 
emulsion will be a definite distance behind the image plane, hence slightly out 
of focus. It may be assumed that this can be fixed up by making the image 
plane of the lens system coincide with the plane of intersection of the rear 
bipack, but this is a false hope. The true image plane is the first diffusing 
surface which intercepts the light rays, hence, under all conditions regardless 



TRIPACKS AND BIPACKS 89 

of the lens optics, the front emulsion constitutes the true plane, and it is most 
desirable that this plane contain as many of the images as possible. During 
its travel through the front element, the diffusion will be sufficient to reduce 
to a considerable extent the enlarging capacities of the negative, but it will 
still be sufficiently sharp to give a usable image in an adjoining emulsion that 
is in optical contact with the first. But if there is a separation between the 
first diffusing surface and the second, the extent and degree of the diffusion 
becomes considerably enhanced. 

This scheme should successfully reduce the blue sensitivity of the basic 
emulsion to balance. The green sensitivity is most probably controlled by 
the conditions under which sensitization takes place. It is rather doubtful 
whether or not, under such conditions, a proper balance is obtained without 
the loss of considerable light efficiency. 

In a patent issued to T. T. Baker (U.S.P. 1867301) the balance in speed 
is obtained in a rather novel manner. At the same time an improvement in 
definition is also claimed in that the red filter negative, which yields the cyan 
image, is placed in front of the pack, followed by the green and the blue in 
that order. The front element is a highly diluted, fine-grained, low iodide- 
content contrasty emulsion, sensitized to the red. These are the conditions 
for optimum action of the sensitizing dye upon the silver halide grain. The 
silver content of the emulsion is one-fifth that of the normal, so that great 
transparency results. The blue sensitivity of this element is further diminished 
by the inclusion of a slight amount of yellow dye. The middle emulsion is a 
similar one, diluted to contain one-fourth of the normal silver content, and 
sensitized with erythrosin. It should have a speed of 350 H & D. The rear 
element is an extremely high-sensitive, color-blind emulsion with a speed rat- 
ing of 850 H&D. 

It is apparent that this is an idealized procedure that is not capable of being 
really produced. First and foremost, there is the disturbing fact that under 
the best of conditions it is not possible to sensitize an emulsion and attain a 
speed in the induced range that is sufficiently greater than the blue speed 
to make it possible to neglect the blue as insignificant. The blue sensitivity 
would have to be less than one-tenth that of the red in the same emulsion, in 
order to be able to discount its effect. If sufficient yellow dye be placed in 
the emulsion to accomplish such a result, there would not be enough blue 
light transmitted to the rear emulsion to yield an image there. Under normal 
working conditions, it is obvious that in this arrangement the red and green 
filter separations would be badly contaminated with blue. If the amount of 
this contamination be known, it will be possible to correct for it by a masking 
procedure. This discussion will be left for a later time. 

In a later disclosure, Mr. Baker goes back to a more normal procedure 
(U.S.P. 1910877). Here he adopts the fundamental idea disclosed by Ives, 
of using the same basic emulsion for all three elements. The front emulsion 
is diluted so that it is 50 per cent as concentrated, as the normal, and it con- 



go HISTORY OF COLOR PHOTOGRAPHY 

tains a basic yellow screening dye, such as auramine. The central emulsion 
is also diluted, but only by 25 per cent, so that it is 75 per cent as concentrated 
as the normal emulsion. It is sensitized with an excess of pinachrome, the 
excess dye acting to prevent any green rays from reaching the rear emulsion. 
The rear emulsion is sensitized to the red, and it is not diluted. In this scheme, 
which appears to be quite reasonable, the balance in speed is obtained by in- 
creasing the transparency of the emulsion by dilution. This also decreases 
the speed, since the number of grains per unit area becomes reduced. The 
balance in contrast scales is maintained within controllable range since the 
same basic emulsion is used for all three elements. E. Sanders Dolgoruki 
also used highly diluted emulsions (Eng. P. 452345). He found that this 
gave a very thin image that was not at all suitable for copy. To offset this 
serious defect, he intensified the silver image. 

The inability to prepare emulsions that have no blue sensitivity makes it 
imperative to place a minus-blue filter somewhere in front of the central emul- 
sion. F. Stolze (Ger. P. 179743) claimed that the emulsion itself is an efficient 
blue absorbent, but this is an overstatement. As much as 25 per cent of the 
blue rays will be transmitted by the normal emulsion. F. E. Ives (U.S.P. 
927244; Eng. P. 7932/08) proposed to deposit a surface layer of yellow dye 
upon the emulsion. This would act as a filter without affecting the blue 
sensitivity of the emulsion itself. He proposed to do this by treating the 
front element with an alcoholic solution of the dye, a procedure that was 
previously disclosed by Husson and Bornot (U.S.P. 922908) in another matter. 
Since alcohol does not wet gelatin, the dye cannot penetrate the gelatin sur- 
face to react with the silver halide grains beneath. 

Theoretically, the tripack should yield speeds that are but slightly lower 
than normal black-and-white emulsions, but due to the scattering properties 
of the photographic emulsion, the loss of light that results from the use of 
two emulsions in front of a third is about of the same order as the loss due to 
light-splitting devices. Efforts to increase the transmission do not seem to 
have been wholly successful, since the only commercial tripack on the market 
at this writing has a speed rating not greater than Weston 3. The rear ele- 
ment that is used with this pack is rated at ten times this amount. To retain 
some of the simplicity of the tripack system, with increased lens speed, there 
was introduced the bipack-plus-one assembly. In this scheme, first intro- 
duced by J. W. Bennetto (Eng. P. 28920/97)' the lens beam is split into two 
parts. One portion is made incident upon a single plate, while the other is 
made incident upon a bipack. Thus three separations are obtained in a camera 
which splits the light into two beams. Of course any of the light-splitting 
devices mentioned in the previous chapters could be used here. We will forego, 
therefore, a discussion of the cameras used to achieve this. It is interesting 
to note that Technicolor utilized a bipack-plus-one camera to obtain its sepa- 
rations. In that procedure, the reflected beam registers the green separation. 
The main beam passes through a magenta filter which removes the green light, 



TRIPACKS AND BIPACKS 91 

then is made incident upon a bipack the front element of which is a color-blind 
emulsion, while the rear element is a panchromatic one. It is even possible 
in this scheme to use an orthochromatic emulsion as the front element, since 
the magenta filter will remove all the green light. Such filters are very effi- 
cient transmitters of both red and blue, so excellent separation quality is 
achieved with a minimum loss of light efficiency. 

It is possible to calculate the speed for a typical bipack, using the fastest 
materials available. The rear element would therefore be a Super Panchro 
Press emulsion with a Weston speed rating for Mazda light of 64. The filter 
factor for the A filter, and incidentally for the red part of the magenta filter 
(Wratten No. 32) is two, therefore the red sensitivity can be assigned a value 
of 32. But the front element will transmit only 40 per cent of the light to the 
rear, so that the effective speed of the rear element is 13. To balance this, 
there must be used a front element, either color-blind or orthochromatic, which 
will have a like sensitivity when exposed through the No. 32 filter. This 
transmits approximately 55 per cent of the total blue, so that the speed of 
the front element should be approximately 24. The Ortho Press emulsion 
would qualify ideally. This emulsion or the Super Ortho Press, when photo- 
graphed behind a K3 filter, has an effective speed of 16. Therefore the lens 
beam must be divided into two equal parts. This will introduce a factor of 2, 
so that it is possible to obtain a speed of Weston 6 by the use of a Super Panchro 
Press and two Super Ortho Press emulsions. This is approximately the same 
speed as would have been obtained by the use of a tripack. But now the cut 
on the blue filter separation is exceptionally good, and there is a considerable 
increase in the quality of the definition. It is seen, of course, that it is possible 
to use other combinations than the one described above, but there will be no 
difference in the light efficiency of the result. Because of the danger that the 
optical system may not give absolute registry, it may be desirable to reflect 
off the blue-filter separation, and use an orthochromatic emulsion in front of 
a pan for the bipack. It will be impossible for the magenta and cyan positives 
to be out of registry in such a case. 

An entirely novel approach to this problem of separations without the use 
of optical systems which divide the light into several beams, is disclosed by 
Ives (U.S.P. 1268847; Eng. P. 112769; Fr. P. 487529). A bipack is used, 
the front element of which is a screen plate that contains yellow and magenta 
lines, and which is coated with an orthochromatic emulsion. The rear element 
is a red-sensitive or panchromatic emulsion. The screen elements are placed 
in front of the pack, so that all light that reaches the emulsions must pass 
through one or the other of the lines. Red light will be transmitted by both, 
but the emulsion coated over the screen elements is not sensitive to red, hence 
it will not register on the front element, but upon the rear, in the form of a 
continuous-tone negative. Blue light will pass only through the magenta line, 
hence these densities will be registered on the front element, immediately 
behind the magenta lines. Green will pass only through the yellow lines, 



92 HISTORY OF COLOR PHOTOGRAPHY 

hence the green densities will be registered on the front element only behind 
the yellow lines. The front element, therefore, yields two primary separations, 
while the rear element yields the third. 

In this form the separations are of little value, since the front element con- 
tains two of them intermingled very closely. But it is a simple matter to 
separate the two primaries. Consider the effect of a print made from the 
front element by means of blue light. The yellow lines in the screen would 
prevent any blue light from penetrating. But free passage to this light would 
be offered by the magenta lines, hence there will be printed only the densities 
that lie behind the magenta lines. If a print were made using green light, 
the magenta lines would prevent the passage of any light, but the yellow lines 
would permit complete transmission. Of course this print must be made upon 
a green-sensitive material. In this way it is possible to physically separate 
the two. A system such as this is really an effective means for increasing the 
speed and the quality of color separations. The transmission of the red rays 
would be quite low, approximately 30 to 35 per cent, due to the absorption 
and scatter in the front emulsion and the slight absorption by the yellow and 
magenta lines. But using very speedy materials that have factors for the A 
filter of 2.0 and speeds of about 100 for the Mazda light, it is possible to 
obtain a pack with an overall speed of 16. A variation of the Ives' bipack 
has been disclosed by G. B. Harrison. He would use a pack in which a color- 
blind emulsion layer is coated on top of a two-line screen beneath which there 
is a panchromatic emulsion (Eng. P. 471586). 

It is rather hard to understand why this system did not immediately re- 
ceive the attention it merited. Here was disclosed an easy and convenient 
method for making separations without the use of a complicated and costly 
optical system, and with a light efficiency that could not be equalled by any 
other method then known. True, the method for making successful line 
screens was not yet developed. Neither was there any great demand by the 
industry for an efficient means to make separations. Colored motion pictures 
had yet to be introduced. Colored illustrations in the magazines were just 
becoming popular, but these were mainly still shots. It was only after the 
success achieved by Technicolor, in the late logo's, that people became con- 
scious that color reproduction was a possibility. With this realization came 
increased pressure for colored illustrations and color photography. But even 
then, no effort seems to have been made to develop the Ives' scheme. 

An optical equivalent of this scheme has been devised by the I.G. (U.S.P. 
2093655; Eng. P. 395124, 454738, 454842, 459 02 7, and 460653). The screen 
is replaced by a lenticular film, coated with an orthochromatic emulsion. Over 
the lens is placed a two-banded filter, dyed yellow and magenta. The theory 
of the lenticular processes cannot be discussed at this time, but is to be left 
for a later chapter. It is merely sufficient to point out that the combination 
of a banded filter in the front nodal plane of the lens system, and a lenticular 
film in the rear focal plane, is an exact optical duplication of a line screen. 



TRIPACKS AND BIPACKS 93 

This disclosure accomplishes exactly what Ives does with his bipack, but with 
much less efficiency, and much more trouble, as will be made evident when 
lenticular film is discussed. 

J. F. Shepherd (Eng. P. 169533) proposed to use a bipack in which the 
front element was orthochromatic and the rear panchromatic. The print from 
the front element was used to give the magenta image while the print from 
the rear was used to give the cyan image. The yellow was made from the 
combined negatives. Obviously this procedure is only a crude compromise 
to the true rendition. It is impossible to obtain yellows, magentas, or purples, 
since these colors would yield equal densities upon both the front and the 
rear negatives. It is to be questioned if this procedure has any advantage 
over a two-color process. A scheme somewhat along these lines was also dis- 
closed by F. Rolan (Ger. P. 406174) and W. R. Whitehorne (U.S.P. 1724445). 

The fact that it is possible to obtain color reproduction by means of two- 
color separation instead of three was first recognized by du Hauron, and 
disclosed in the paper and patent noted at the beginning of this chapter. He 
was quickly followed by a host of others. A. Gurtner (Eng. P. 7924/03; 
U.S.P. 730454), used a front element that was sensitive only to the blue, and 
a rear element that was sensitive up to but not including the red. He was 
the first person to suggest that the two films or plates be placed emulsion to 
emulsion, and have their edges pasted together to form a single pack. Finnigan 
and Rodgers (Eng. P. 140349) would dispense with a filter in the front element 
or before the rear, claiming that the emulsion itself acted as a filter. This is 
not quite accurate, as it merely acts as a filter with a density of 0.60, a value 
quite insufficient for practical results. 

To equalize the sensitivity of the rear emulsion with that of the front, 
A. F. Cheron (Fr. P. 444599) placed a filter in front of the lens that contained 
an inner circle that was clear, and an outer circle that was orange-colored. 
The front element was therefore exposed only by the cone of light that was 
transmitted by the inner clear area, while the rear element received this light 
and also the light transmitted by the orange-colored outer circle. A. B. Klein 
and T. S. Wilding (Eng. P. 449591) suggest a bipack that could be used for 
two- or three-color separations, in which the front element is red-sensitive, 
and the rear element is orthochromatic. The front element records, therefore, 
red and blue, while the rear element records the green. When used with a 
yellow filter in front of the lens, there is eliminated the necessity of having the 
filter incorporated into the front element. But such a system, when used in 
two-color work, would record all the pure blues and purples as blacks and 
grays. 



CHAPTER 9 
MONOPACKS 



Th, 



lHE most modern procedure for the elimination of complicated camera 
devices to achieve color reproduction lies in the use of monopacks. These 
substances are multi-layered films in which each layer contains an emulsion 
sensitized for a single primary color. The different layers are not separable, 
but must be treated as a single unit. This means that a new technique must 
be used in order that the images in the different emulsion layers be separated 
from each other. The monopack film is merely an integral tripack. 

The multiple coating of emulsion layers one on top of another has been 
known for a considerable length of time. H. Kuhn (Eng. P. 6921/91) dis- 
closed such a system in 1891, although he made no attempt to utilize it for 
the purposes of color photography. A mount is made waterproof, and is 
then coated with an emulsion of barium sulphate in gelatin. On this is coated 
a series of sensitive gelatin or collodion silver bromide emulsions (print-out), 
mixed with aniline colors. One emulsion is coated on top of the other after 
the first has dried. Each layer is dyed a different color. 

From this description it is not such a stroke of genius to make the system 
applicable to color photography. The dyes in each of the layers can be so 
chosen that they act as filters or color sensitizers. Therefore it is possible 
to coat one emulsion in its native state so that it will be blue-sensitive only. 
A yellow dye could be incorporated in this layer. The next emulsion layer 
is dyed with a green sensitizer. The yellow dye in the front layer can be 
used to prevent any blue light from reaching this or the third layer. This 
last is dyed with a red sensitizing dye. In 1891 the powerful carbocyanine 
and cyanine sensitizers were not known, so that the designation of aniline 
dyes may be taken to mean sensitizing as well as filter dyes. The specifica- 
tion does not disclose why the dyes were put into the emulsion, nor why the 
multilayered paper was needed. But it does disclose the preparation of a 
monopack. 

The first specific disclosure of the use of a monopack for the purposes of 
color reproduction came from K. Schinzel {Phot. Woch., 1905; Brit. J. Phot., 
Vol. 52 (1905) p. 608; Aust. P. 42478/08). The plate is coated with three 
silver-bromide emulsions colored complementary to their spectral sensitivities. 
The individual layers must be dyed with colors that do not bleed, and which 
are insoluble in water. The top layer is colored yellow and is sensitive to the 
blue. The second layer, colored cyan, is sensitized for the red. The bottom 

94 



MONOPACKS 95 

layer is colored magenta and is sensitized for the green. This is as clear and 
as definite a disclosure of a multilayered monopack for color reproduction as 
it is possible to make. 

The pack is exposed, developed, and fixed in the ordinary manner. In this 
condition each layer contains a silver image of a single primary, imbedded 
in a layer of a complementary-colored gelatin. Upon treatment with two per 
cent peroxide, it is claimed that the dye in the immediate vicinity of the 
silver image will be bleached, and in an amount directly proportional to the 
image density. Therefore there remains a dye image in each layer which is 
the negative of the silver image. We will leave a critical discussion of the 
processing technique to a later chapter where we shall discuss the dye-bleach 
process of color reproduction. Here it is sufficient to point out that in 1905 
there was disclosed completely the use of a Kuhn monopack for color repro- 
duction purposes, thus anticipating Mannes and Godowsky, Troland, Gaspar, 
and a host of other inventors, who have made this system the best answer 
so far (1943) to the problem of original exposure. 

Needless to say neither the state of the photographic art, nor that of the 
dye industry was sufficiently advanced to allow the Schinzel disclosure to be 
put to practical use. It remained for Dr. Bela Gaspar to accomplish this 
result. But the interest which it aroused was immediate, and the discussion 
that followed amplified the procedure.. Neuhauss {Phot, Rund., Vol. 19 (1905), 
p. 239) pointed out that the use of peroxide was not advisable as this chemical 
would not discriminate very carefully between the dye lying adjacent to a 
silver deposit, and the dye that lay in non-image portions. Schinzel admitted 
the fallacy, and suggested the admixture of the emulsion with colorless sub- 
stances which would afterward give rise to color formation (Chem. Ztg., Vol. 32 
(1908) p. 665; Brit. J. Phot., Vol. 55 (1908) Col. Supp., Vol. 2, p. 61). 

Another suggested use of a monopack for color reproduction came in 1910, 
from F. Sforza (Phot. Coul., Vol. 5 (1910) p. 209; Brit. J. Phot., Vol. 57 (1910), 
Col. Supp., Vol. 3, p. ^). In this disclosure, the three layers were to be dyed 
with the primary colors that were mordanted to the gelatin. The dyes should 
have the added property of being catalytically destroyed in the presence of 
the silver. Somewhat similar ideas were expressed by R. Luther (Phot. Rund., 
Vol. 25 (1911) p. 1). 

From this, it is very evident that monopacks were well established by 1924 
when Mannes and Godowsky were issued their first patent (U.S.P. 1516824). 
In this early patent there is disclosed a general two-layered monopack wherein 
the film base was first coated with a fast red-sensitive emulsion, then topped 
with a slower orthochromatic emulsion that contained a yellow dye dispersed 
throughout its depth. This prevented the blue rays from affecting the red- 
sensitive layer. The green rays would act only upon the top emulsion layer, 
since the bottom layer was not sensitive to the green. A two-color separation 
was achieved. If the two emulsions were correctly balanced, a single exposure 
yielded two equivalent latent images, one in each of the layers. By develop- 



q6 HISTORY OF COLOR PHOTOGRAPHY 

ment the latent images were converted into metallic silver. The pack was 
then fixed and washed. Treatment with ferricyanide converted the silver into 
reducible silver salts. A further treatment with a developer compounded to 
diffuse very slowly into the depth of the emulsion, made it possible to confine 
the developer action to the upper layer only. This achieved the separation 
of the two images, for one was now in the form of a metallic silver image 
that could be chemically or dye toned to any desired color, while the other 
was in the form of a silver salt which could be treated independently to form 
a different color. A slight variation of this was contained in a later disclosure 
by the same men (U.S.P. 1659148). 

At about the same time a large number of other inventors entered the field 
of monopack photography. Since the multiple coating of emulsion layers of- 
fered very little chance for invention, most of the disclosures differed among 
themselves only by the procedures which were adopted to separate the images 
in the different layers. J. F. Leventhal (U.S.P. 1697 194) treated the pack, 
after development and fixation, with a chemical that retarded the action of a 
bleach. In that manner he was able to bleach the image in one layer without 
affecting the other. This left the film with one silver and one silver-salt image, 
which could be differentially dye or chemically toned. 

The next advance came in 193 1 in a series of patents issued to Dr. L. T. 
Troland, then director of research of the Technicolor Motion Picture Cor- 
poration (U.S.P. 1808584, 1928709, and 1993576, reissue 18680; Eng. P. 370908 
and 382320). These disclosed several methods for the preparation of the mono- 
pack, several uses for the material, and several methods whereby it became 
possible to separate the images. There is no novelty in the obvious coating 
methods for the formation of the monopack, so we shall discuss only the novel 
forms he disclosed. Dr. Troland recognized that it is very difficult to separate 
the images in a three-layered monopack, so he proposed to use a two-layered 
pack sensitized to yield the blue and the red separations. This monopack he 
then used as the front element of a bipack, the rear component of which was 
a green-sensitive emulsion. This is merely a generalization of the Ives bipack 
disclosure where a screen plate is used as the front element. 

Two methods were described for the formation of the monopack, which 
forms the front element. One is the obvious method previously disclosed by 
Kuhn, Schinzel, Sforza, etc., which consisted in physically coating properly 
sensitized emulsions, one on top of the other. There is very little to discuss 
here except that Dr. Troland preferred to have a separate filter layer, dyed 
yellow, between the two emulsion layers. The second method was to treat a 
color-blind, blue-sensitive plate or film, with a sensitizing solution under the 
conditions of controlled diffusion. This can be done quite easily if the sensi- 
tizing bath be compounded with a high alcoholic content. Under these con- 
ditions the colloidally dispersed sensitizing dye (pinacyanol) does not penetrate 
deeply into the depth of the emulsion. After drying to achieve thorough sensi- 
tization, the plate is bathed in a water solution of a yellow dye. This penetrates 



MONOPACKS 97 

more deeply than the sensitizing dye. By this method the emulsion thickness 
of an ordinary negative material is divided into a lower blue-sensitive zone, 
and an upper red-sensitive one. The yellow dye, dispersed much more deeply 
than the sensitizer, prevents the blue rays from penetrating into the zone of 
red sensitivity, hence there is no color contamination. This bathing method 
yields a product which is quite unstable, but which otherwise is altogether 
suitable. 

Dr. Troland was not the first person to disclose that stratification can be 
achieved by bathing the emulsion with the colloidal sensitizers. Indeed it was 
a well known and established fact that unless extra precautions were taken when 
sensitizing an emulsion by bathing, only a superficial layer of the coating 
would be affected. But a very clear statement of this phenomenon, and a 
clear disclosure as to the effect that an increase in alcohol concentration would 
have, is contained in a sensitizing patent by G. Selle (Eng. P. 12516/99) who 
wrote: "To achieve my object I use a sensitizing bath, for instance, a solu- 
tion of 0.002 per cent of cyanine blue and erythrosin in a fluid composed of 
60 per cent water and 40 per cent alochol. By this means the red dye (ery- 
throsin), which is more soluble in the water, is carried into the film, while the 
blue dye (cyanine blue), which is more soluble in alcohol, remains substan- 
tially on the surface of the plate. But other sensitizers having the same 
properties or results may be used." Pinacyanol or pinacyanol blue are red- 
sensitizing dyes that have the solubility properties of cyanine blue. Troland, 
therefore, did not discover the art of surface sensitization or stratification. 
But he appears to have been the first one to realize that this stratified emulsion 
could replace the screen plate in an Ives' bipack. 

The Troland patent was issued in 1931, but it was applied for in 1921. For 
this reason Dr. Troland was given priority to Mannes and Godowsky. Since 
the Eastman Kodak Company controls both these disclosures, the question 
of priority is merely academic and economic. The Troland patent is valid 
to 1948, while the Mannes and Godowsky patent expired in 1941. 

Independent of the above, Emil Wolff-Heide disclosed a similar scheme for 
surface stratification (Eng. P. 340278) . He based his work upon the well known 
fact that pinacyanol was a colloidally dispersed dye when dissolved in a water- 
alcohol mixture, and that gelatin is an excellent semi-permeable membrane 
for colloids. Von Httbl ("Die Orthochromatische Photographie," (1920) p. 69) 
has the following to say in this respect: "Such solutions are colloidal and tend 
to flocculate; if one filters it, the dye remains on the filter paper and the 
solution comes through colorless. . . . (During sensitization) the dye sepa- 
rates out for the most part upon the surface of the film, and only a colorless 
solution, which does not sensitize well, penetrates into the innermost parts 
of the film." Mr. Wolff-Heide found it sufficient to bathe a film in a pinacyanol 
solution, together with filter yellow, rose bengal, and pinachrome. This is 
but slightly different from the Troland disclosure. Since his patent was issued 
several years prior to that of Dr. Troland, he could not have been aware of 



98 HISTORY OF COLOR PHOTOGRAPHY 

the information contained in the latter. H. von Fraunhofer patented essen- 
tially the same (U.S.P. 2030903 and 2030904). In the second patent it is 
suggested that the sensitizing solution be placed upon the subbed film-base 
prior to the coating. The sensitization under those conditions would take 
place at the bottom of the layer, so that the zone of red sensitivity would be 
that part of the emulsion layer that lies adjacent to the base. The two silver 
images were separated by treatment with a uranium toning solution. The 
heavy uranium ion is ideally suited for use in controlled diffusion processes. 
The time of treatment is so arranged that only the upper silver image becomes 
toned. The lower image could be converted into a blue by any other method 
that would not affect the uranium-toned image. 

A more detailed description of a sensitizing technique, which will yield a 
stratified emulsion, is contained in United States patent 2047022 issued to 
J. S. Friedman, and assigned to the Omnichrome Corporation. The sensi- 
tizing bath is concocted as follows: 

Rapid filter yellow 1.0 part 

Water 600 parts 

Pinacyanol 0.015 part 

Alcohol 400 parts 

This bath contains 40 per cent alcohol. The time of sensitization is approxi- 
mately two minutes at a temperature that is as close to 40 to 45 F as is possible. 
It is also permissible to pre-wet the emulsion by immersion in a water solution 
made slightly alkaline by the addition of borax. This will swell the gelatin 
to a considerable extent and it will therefore give a greater degree of control 
to the diffusion of the sensitizing agents into the emulsion. A somewhat more 
effective reaction between silver halide and dye is also obtained when the 
emulsion is thus pre-wetted. 

Another method whereby a stratified emulsion is obtained, is disclosed in a 
patent issued to J. S. Friedman and A. Bruck (U.S.P. 2175836) and assigned 
to Color Processes Inc. Here advantage is taken of the fact that an alcoholic 
solution of a dye will not penetrate into the depth of the emulsion, but will 
leave a layer of dye deposited uniformly over the surface. Ives utilized this 
phenomenon to deposit a filter layer of dye upon the surface of an emulsion. 
In 1917 Koenig (Phot. Korr., Vol. 55 (1918), p. 22; Phot. Rund., Vol. 54 (1917), 
p. 257) recommended this as a fool-proof method for sensitization. But he 
did not recognize that with a slight modification, the procedure could be used 
to give color separations. 

A blind emulsion is first bathed in an alcoholic solution of the red sensitizer 
containing 0.030 gram of dye to the liter of alcohol. Time or temperature 
plays no part whatsoever in this step, since all that happens is that a thin 
layer of the red sensitizer is deposited upon the surface. Only a sufficient 
quantity of dye is deposited to react with the upper half of the emulsion 
layer. 



MONOPACKS 99 

After drying, the plate with the surface coating of dye is immersed in a 
solution that contains 20 per cent alcohol and 0.015 per cent auramine, pre- 
viously purified by recrystallization from alcohol. As the water-alcohol mix- 
ture becomes absorbed by the gelatin, it dissolves the sensitizing dye, and 
carries it to the silver halide grains. Sensitization takes place until all the 
dye has been used up. Since only sufficient dye is deposited to react with 
the upper half of the emulsion layer, stratification of the emulsion results. 
The yellow dye prevents the blue rays from reaching the zone of red sensitivity 
if the exposure is made through the base. The Friedman monopacks were 
designed to be used as the front elements of a bipack. The rear component 
of the combination, being an orthochromatic emulsion, yields an image of 
the green primary. 

In order to process these monopacks, due regard must be given to the fact 
that the developer will start its action first upon the latent image present in 
the upper portion of the emulsion, and that by the time the developer reaches 
the lower layer, it is considerably weaker. It is desired, if the two images 
are to be developed in a single solution, to compound a developer that diffuses 
very rapidly into the gelatin. This can be done if the carbonate content or 
the alkalinity is very high. But this has the drawback that considerable fog 
will be developed. Dr. Troland recommended a developer of this type, com- 
pounded by adding 10 parts of carbonate to one part of hydroquinone. Another 
drawback is that such a developer is very contrasty, and therefore hardly suit- 
able for accurate photographic reproduction. Approximately the same result, 
but with much more accurate control of the contrast scale, could be obtained 
if the development were carried out in two stages. First the pack could be 
bathed for ten or fifteen minutes in a solution that contains all the ingredients 
of the developer, with the exception of carbonate, thus : 



Metol 


5.0 parts 


Sulphite 


30 parts 


Bromide 


0.5 part 


Water to 


1000 parts 



In this solution the development action will just start in approximately ten 
or fifteen minutes. This indicates that this length of time will be sufficient 
to bring about complete saturation of the emulsion with developer solution, 
and, more important, to allow adsorption equilibrium to be established be- 
tween latent image and developing ions. From this solution, and with no 
intermediate wash, the pack is treated with a 10 per cent solution of sodium 
carbonate, containing one gram of bromide per liter. Development will be 
complete in approximately three to five minutes. Because the carbonate is 
so concentrated, there will be no appreciable diminution in its strength on its 
passage through the upper image layer. 

It is possible to take advantage of the fact that two separate layers are to 
be developed. Here, again, a two-stage development is in order, but this time 



ioo HISTORY OF COLOR PHOTOGRAPHY 

each stage utilizes a solution capable of yielding a developed image. The first 
stage will use what is known as a surface developer, one of the fine-grain low- 
potential developing solutions that is achieving so much current popularity. 
One such developer could be concocted as follows: 



Metol 


z\ parts 


Sulphite 


ioo parts 


Tri-ethanolamine 


i part 


Bromide 


\ part 


Water to 


iooo parts 



The high sulphite and low alkalinity retards the diffusion of the developer 
to the lower levels. This can be further improved by the addition of 200 parts 
of desiccated sodium sulphate, or 15 parts of chrome alum. If the alum is 
used, it would be necessary to add from five to fifteen parts of neutral sodium 
citrate to prevent the precipitation of chromium hydroxide. 

After the plate has been developed sufficiently in this solution, usually about 
five or ten minutes, it is washed for ten or twenty minutes in running water 
to remove all developer and developer oxidation products. It is then de- 
veloped in a developer with a high appearance time (low Watkins factor), a 
high diffusion constant, and potassium iodide. One such developer would be 



Metol 


ipart 


Sulphite 


30 parts 


Hydroquinone 


5 parts 


Carbonate 


30 parts 


Potassium iodide 


5 parts 


Water to 


1000 parts 



This is a typical hydroquinone developer which has a very high appearance 
time. During this period the iodide reacts with the silver bromide in the upper 
layer, and converts it to undevelopable silver iodide. The main reaction is 
thus forced to take place below in the bottom layer. It may be advisable to 
interpose an alcoholic iodide bath between the two development stages. This 
will convert the silver halides in the upper strata to silver iodide. A thorough 
wash removes the excess iodide, after which the lower layer can be developed 
at will in a normally compounded developer. 

The Friedman disclosures do not utilize a fixation reaction whereby the un- 
reduced silver halides become dissolved out by the action of hypo. Instead, 
after the development is complete, the pack is again washed thoroughly, then 
subjected to the action of one per cent solution of potassium iodide, for a 
period of fifteen to thirty minutes. After a thorough wash, the pack is dried. 
In this condition the pack contains two black-and-white silver images imbedded 
in a yellowish-white, opaque, silver-iodide layer. When viewed from the emul- 
sion side, the only one of the two images that will be seen is the image of the 
red densities. The same is .true if the pack be viewed from the base side, but 



MONOPACKS 101 

this time it will be the blue densities that will be seen. To separate the two 
images, it is sufficient to copy each, side of the pack upon a different plate or 
film. In making these copies or "take-offs" it is desirable to use an ortho- 
chromatic process emulsion, and a K3 filter to offset the yellow stain of the 
silver iodide. 

The reflection copies of the images on the pack will be positives. In some 
processes positives are required from which to print on to the coloring material. 
But in most cases it is desirable to have the separations in the form of nega- 
tives. It is possible, of course, to reverse the copies. This can be done quite 
easily. After the copy has been exposed it is developed in a developer to which 
is added ten or fifteen grams of hypo per liter. A thorough wash removes 
excess developer. Treatment with acid dichromate or acid permanganate 
etches out the silver image. By this treatment the silver is oxidized to ionic 
state, in which condition it has a great affinity for gelatin. The silver-gelatin 
complex must be destroyed and the silver ions removed before the next step 
is taken, otherwise this will be reduced to metallic silver again in the second 
development. A rinse in weak ammonia, or a bath of one per cent sodium 
sulphite usually accomplishes this. The washed plate is next exposed to light, 
or treated with thiourea or methylene-blue solutions which act exactly like 
light. It is then developed in a second developer that does not contain hypo. 
Or, the plate after the treatment with sulphite can be treated with sodium 
sulphide, sodium hydrosulphite or any other reagent which will convert it to 
a black. 

The original plate can be processed so that it yields a positive rather than 
a negative. To this end it is possible, after the development of the two latent 
images is completed, to treat the plate with a strong sodium sulphide solu- 
tion. This will convert the entire plate into a deep black, in which there is 
imbedded two silver images. The oxidation of silver to an insoluble salt pro- 
ceeds much more rapidly, and at a considerably lower potential, than the 
oxidation of silver sulphide. It is therefore possible to find a mild oxidizing 
agent which will attack the silver and convert it into a white insoluble salt, 
and which will leave the silver sulphide intact. Mercuric chloride, cupric 
chloride, and ferricyanide solutions are substances that have this property. 
By this treatment there will be formed a white negative image imbedded in 
a black background, and this is identical with the presence of a black positive 
image imbedded in a white background. 

Another possible routine would take advantage of the mordant power of 
silver iodide for basic dyes. After the monopack has been completely proc- 
essed so that it contains two silver images imbedded in a background of silver 
iodide, the plate can be treated with a black basic dye, or with a mixture of 
methylene blue, rhodamine B, and auramine. The silver iodide will absorb 
the dyes deeply, especially if the iodide content of the final bath be increased 
to five or ten per cent. This treatment is to be followed by conversion of the 
silver to a white insoluble salt by treatment with mercuric, cupric, or ferri- 






102 HISTORY OF COLOR PHOTOGRAPHY 

cyanide ions, in the presence of halides or other cations whose silver salts are 
white and insoluble. 

A further possibility would lie in the fact that soft gelatin is a strong absorber 
of certain dyes. After development of the monopack it can be fixed in alum- 
free hypo, thoroughly washed, and then treated with a bromoil bleach. This 
will harden the gelatin immediately surrounding the image, in which condi- 
tion it will no longer absorb certain dyes like platinum black, or the pinatype 
dyes, while the rest of the gelatin will be very receptive of them. A uranium 
toning solution will affect the image gelatin in a similar fashion. Here again 
there will result a white negative image imbedded in a black background, or 
a black positive imbedded in a white background. 

Dr. Bela Gaspar has disclosed a bipack which yields three color separations. 
The front element (Eng. P. 448161 and 450685) of this bipack was a monopack 
formed by coating two emulsion layers one on top of the other. What makes 
this pack differ from the one disclosed by Troland is that the layers contain 
dyes that are complementary to their sensitivity, and which by treatment 
with special chemicals become decolorized in direct proportion to the intensity 
of silver deposit. The preparation and chemistry of these solutions will be 
discussed in a later chapter, devoted to the Gaspar disclosures. The I.G. com- 
pany also disclosed a bipack (Fr. P. 836173). The front element is a two- 
layered monopack that contains substances that react with the oxidation prod- 
ucts of the developer to form insoluble dyes. These substances are called 
"couplers," and the process of development in which they are used is called 
"color-development." Because of the importance of this technique it will be 
discussed in a later chapter devoted exclusively to it. The tendency has been 
to utilize color-development almost exclusively with monopacks, so much so 
that the two appear to be bound in holy wedlock. This is a marriage of con- 
venience, and it is not at all essential. 

As was indicated above, monopacks are probably the most popular of all 
systems of photography at the present time, and this field is receiving con- 
siderable attention from all serious workers. Almost every film manufacturer 
has taken out patent upon patent to protect some supposedly novel feature 
in assembly or processing technique. But, truth to tell, there is very little 
real novelty in most of the disclosures. Gaspar (Eng. P. 421534; Fr. P. 753061) 
follows the lead of the I.G. company and adds substances to the emulsion that 
can be converted into colors. The duPont company would process the monopack 
film with color developers (U.S.P. 2133937 and 2140540; Eng. P. 497463 and 
497698). In United States patent 2 1666 17 Sease and Weber, of that company, 
disclose a scheme to treat the upper layer of the pack with concentrated 
solution of hypo (50 to 90 per cent). In English patent 505861, assigned to 
the duPont company, there is disclosed the following technique: A record is 
made of the combined images in the pack. These are then bleached to silver 
salts, and only the upper image is redeveloped. From these, the other image 
can be obtained by a system of masking. 



' 



MONOPACKS 103 

J. H. Reindorp (U.S.P. 2153698; Eng. P. 465090, 467005 and 467380) dis- 
closes another method for processing each of the two layers of a monopack, 
individually. The method utilizes a controlled diffusion of an iodide solution 
to convert the upper emulsion layer or stratum into silver iodide. It is a known 
fact that when a silver chloride grain is converted into bromide or iodide, it 
does not lose its latent image. Therefore, the latent image existing in the 
upper stratum is not destroyed, although it will not be developable by means 
of the ordinary developer, but will require solutions with extremely high po- 
tentials. And upon this fact depends the entire procedure. 

After exposure, the monopack is bathed in Renwick's solution, which is as 
follows: 

Potassium iodide 10 parts 

Hypo 15 parts 

Sodium sulphite 20 parts 

Water to 1000 parts 

The time of treatment is adjusted so that the action is limited to the upper 
stratum only. The plate is washed thoroughly, then developed in 

Diethyl-paraphenylene-diamine hydrochloride J part 

Sodium sulphite 5 parts 

Potassium carbonate 20 parts 

Potassium bromide 0.1 part 

Water to 1000 parts 

To this is added a half gram of para-nitro-benzyl-cyanide dissolved in 50 cc 
alcohol. In this solution the latent image in the non-iodized layer (the lower 
one) will be converted into silver. But at the same time there will be deposited 
in situ with the silver, and in an amount directly proportional to it, a magenta 
dye which is insoluble in alkaline media, but very unstable in acid. 

The upper iodized latent image can now be developed by means of alkaline 
amidol, or by the following: 

Metol 5 parts 

Potassium sulphite 75 parts 

Sodium carbonate 25 parts 

Potassium bromide J part 

Water to 1000 parts 

The unreduced silver halides are removed by treatment with hypo. The 
upper silver image is bleached with 2 per cent ferricyanide, 2 per cent am- 
monia, and sufficient alcohol and glycerin to prevent the penetration of the 
bleach to the lower depths of the emulsion. The newly formed silver salt is 
next developed with a solution that deposits a yellow dye beside the silver 
image. At this point the monopack film contains two silver and two dye 
images, the magenta lying in the bottom, and the yellow in the top layer. 
The silver is removed by the action of Farmer's reducer made slightly alkaline 
to prevent any action on the dyes. This leaves only the pure dye images. 



104 HISTORY OF COLOR PHOTOGRAPHY 

These can be separated by filter action, since the magenta will absorb green 
light, and the yellow will absorb blue light. The yellow developer is formed 
by adding solution B to A, then adding 10 cc of a z\ per cent solution of 
alpha naphthol in alcohol, to every ico cc of the mixture 

A, 1:2:6 amino naphthol sulphonic acid 25 parts 

Sodium sulphite 50 parts 

Water to 1000 parts 

J3. Potassium carbonate 75 parts 

Water to 1000 parts 

This patent has been assigned to the True Colour Film Co., an organiza- 
tion which has done considerable work in the field of monopacks and color 
development. Other patents issued to them which are concerned with mono- 
packs and their processing are United States patents 2137785 and 2163325; 
English patents 453674, 465765, 480251, 480287, 480291, 481274, 483020, 
483035, 49 8 749> 49 8 762 and 505099. 

Cinecolor is another firm actively engaged in monopack procedure. A film 
containing two images is treated with a mordanting solution whose action is 
limited to the upper image only (U.S.P. 2009689; Eng. P. 447412 and 459234). 
The action of the bleach is arrested by treatment with bisulphite. In this 
manner the two images are separated, for one is now in the form of a mordant 
image which can be dyed to any desired color, while the other (lower) is still 
in the form of a silver image which can be toned to yield the complementary 
color independently of the upper one. In another disclosure (Eng. P. 473993) 
provision is specifically made for the reflection printing of the two images. 
To this end there is interposed between the two emulsions a layer containing 
a yellow dye and some other chemicals that can later be treated to form an 
opaque, but highly reflective, substance. This intermediate layer can contain 
zinc hydroxide. After development, fixation, and washing, the film is treated 
with a solution containing sodium sulphide. This will convert the zinc hy- 
droxide into zinc sulphide, a white substance that is a good reflector. Or 
the intermediate layer can be made to contain silver iodide. This will be 
sufficiently opaque to act as a reflector. 

Comstock (U.S.P. 1956274) would process the monopack with a developer 
that yields a reflective image. Such a developer can be concocted as follows: 



Hydroquinone 


5 parts 


Sodium sulphite 


20 parts 


Ammonium carbonate 


60 parts 


Water to 


1000 parts 



The developed silver image is white. After fixation, the two images can be 
printed by reflection from the opposite side of the film. 

Following the acquisition by the Eastman Kodak Company of the Mannes 
and Godowsky disclosures, a whole series of patents dealing with monopacks 



MONOPACKS 105 

and methods of processing them, were issued various members of the staff. 
Seymour (U.S.P. 1897866) would sensitize the different layers with dyes that 
are not destroyed by the action of acid dichromate. After exposure and de- 
velopment, the silver is removed by the action of acid dichromate. The sensi- 
tivity of the remaining silver halide is restored by a bath in 0.5 per cent sodium 
sulphite. At this point the film is exposed to red light. This will affect only 
the grains that are sensitive to the red. Development is accomplished by 
means of a solution that deposits a dye together with the silver. The remaining 
silver salts are then exposed to white light and developed in a solution that 
yields a color different from the first. 

In another disclosure (U.S.P. 1900870) the monopack containing two layers, 
one of which is sensitized to the red with a stable dye, is exposed and then de- 
veloped for seven minutes in D-16, which is concocted as follows: 



Metol 


0.31 part 


Sodium sulphite 


59.6 parts 


Hydroquinone 


6 parts 


Sodium carbonate 


18.7 parts 


Potassium bromide 


0.86 part 


Citric acid 


0.68 part 


Potassium metabisulphite 


1.5 parts 


Water to 


1000 parts 



After a ten-minute wash to remove the excess developer, the plate is exposed 
to red light and developed in the following: 



Sodium carbonate 


20 parts 


Sodium sulphite 


2.5 parts 


Diethyl-para-phenylene- 




diamine hydrochloride 


5 parts 


5 % alcoholic di-brom-ortho- 




cresol 


100 parts 


Water to 


1000 parts 



In this bath the exposed silver halide grains will be reduced to silver and at 
the same time there will be deposited in situ and in exact proportion to the 
silver, a cyan dye. Hence there will be formed in the layer containing the 
red-density record, a cyan positive image of this record. 

The other layer is then exposed to white light, and developed in a magenta 
dye-coupling developer made by adding solution B to A, and diluting to one 
liter 

A . Sodium carbonate 40 parts 

Sodium sulphite 5 parts 

Diethyl-para-phenylene- 

diamine hydrochloride 5 parts 

Potassium bromide 1 part 

Water to 400 parts 



io6 HISTORY OF COLOR PHOTOGRAPHY 



B. Alcohol 


200 parts 


Cyanaceto phenone 


5 parts 


Ethyl-alpha-chlor-aceto- 




aceticester 


1 part 



The silver can be removed by treatment with Farmer's reducer, leaving pure 
dye images that can easily be separated by filters. 

Capstaff (U.S.P. 1954346) simplified this treatment. The monopack, one 
of whose layers is sensitized by means of a dye that is stable to acid dichromate, 
is exposed, developed, washed, and then treated with acid dichromate to re- 
move the silver images. It is next exposed to red light (if the stable dye is 
the red sensitizer) and developed with a cyan dye-coupling developer, such 
as the one given above. After this development, there is present a silver-plus- 
cyan dye image in one layer, and a silver halide image in the other. The silver 
halide image can be treated in a number of ways to convert it to a color dif- 
ferent from that of the first, and in that way achieve an optical separation of 
the two images. 

Burwell (U.S.P. 1966330) varied the technique somewhat. The monopack 
film is developed after exposure, in a cyan color-developer such as the one 
disclosed in the Seymour patent. Fixation in neutral hypo, is followed by a 
thorough wash. At this stage the monopack contains two silver-plus-cyan dye 
images in the two layers. The image in the upper layer is destroyed by 
treatment with acid-ferricyanide-bromide, which converts the silver to silver 
bromide, and destroys the dye. It is then redeveloped in a magenta color- 
developer, or toned red by means oi a uranium toner. 

These ideas bring to mind another possible method for separating the. two 
images in a two-layered monopack. After exposure and development, the 
excess developer could be washed out very carefully, and the pack then bathed 
in a neutral solution of one per cent auramine. An exposure with blue light 
will not penetrate very far into the depth of the dyed emulsion, so that it 
will be possible to expose but one layer at any one time. If the first exposure 
will be upon the side carrying the red-sensitive layer, the development can 
be made with a cyan developer. Now again we have a cyan plus metallic 
silver positive image in one layer, and a bromide image in the other that 
can be converted into a magenta by any one of many methods. 

It is not even necessary to use color developers at this point, for after a 
single layer has been exposed, there exist already two distinct and difleren- 
tiable images, one a latent and developable silver halide image, and the other, 
an undevelopable silver halide image. The development can proceed in a 
normal manner so that after this treatment there will be formed two images: 
one a silver positive image, the other a silver halide positive image. The last 
can be converted, by treatment with strong iodide or thiourea solution (cf. 
chapter on Dye Toning), into a mordant capable of absorbing basic dyes. The 
silver image can then be converted into an iron blue or copper, uranium, or 
nickel-dimethyl-glyoxime red. 



MONOPACKS 107 

Murray and Spencer in a series of disclosures also describe a monopack and 
a system for its processing to yield diflerentiable images. These disclosures 
(U.S.P. 2140847 and 2163325; Eng. P. 440422, 470074 and 489299) are not 
sufficiently novel to merit critical discussion. 

A novel type of monopack is disclosed in a series of patents issued to D.K. 
Allison and L. M. Dieterich and assigned to Detracolor (U.S.P. 2005790, 
2014606, 2034220, 2034230, 2036994, 2151065 and 2161735; Eng. P. 462232). 
The different layers of the monopack film contain the leuco bases of the more 
stable diphenyl methane dyes. These bases are unaffected by treatment with 
developers and hypo solutions, and they have no effect upon the sensitivity 
of the emulsions. The exposed, developed, and fixed monopack is finally 
treated with a bleach such as 



Potassium ferricyanide 


5 parts 


Chromic acid 


1 part 


Copper sulphate 


2 parts 


Hydrochloric acid 


5 parts 


Water to 


1000 parts 



This oxidizes the leuco bases to the dyes and at the same time converts the 
silver image to a silver-copper complex that acts as a mordant. Bromine or 
chlorine water can also be used. It appears that the chemistry involved in 
the bleach solution may be a little weak. Unless a solution which contains 
both copper, and ferricyanide ions is heavily loaded with citrate, oxalate or 
other organic polycarboxy or polyhydroxy acids, a precipitate of cupric ferri- 
cyanide is formed. But this is a minor defect in the disclosure, since it is 
possible to remedy it by the inclusion of the proper stabilizing ingredients. 






CHAPTER 10 
KODACHROME AND KODACOLOR 



A 



SUCCESSFUL monopack film was introduced commercially in the 
spring of 1935 by the Eastman Kodak Company. It bore the trade name, 
Kodachrome. This name is a little confusing since it was previously applied 
to another Eastman product, which had been abandoned for some time prior 
to 1935, and in which transparencies in color were made by a pinatype pro- 
cedure. This will be discussed in greater detail in the section that deals with 
the pinatype processes. We will follow Eastman in restricting the term 
Kodachrome to the monopack. The new process was an overnight success, 
and it helped make the photographic amateur color-conscious to such a degree 
that great pressure soon began to be felt in the entire industry to make color 
available to all, in a cheap and simple manner. Simplicity was perhaps a 
greater requirement than cost. For this reason, this discussion of the Koda- 
chrome process will be rather detailed. A source rather high in the Eastman 
councils (E. R. Davies, of the British Kodak Research Laboratories) has ex- 
pressed the opinion that the process can be ascribed to two men, L. D. Mannes, 
and L. Godowsky, members of the Eastman research staff. A discussion of 
the process becomes therefore, a discussion of the patents issued to these men. 
Of course not all the disclosures made by them are incorporated in the Koda- 
chrome technique, but it is interesting to note exactly how the procedure 
evolved from a mere concept to a finished product. 

The first application which Mannes and Godowsky made, and which termi- 
nated in the grant of a patent, bears the application date October 4, 1921. It 
is recalled that Mr. Mannes was at Harvard during the years 1917-20, so it is 
apparent that serious interest in color came at a very early period in his life. 
This first patent (U.S.P. 1538996) does not disclose anything really new or 
very startling. It relates to the making of a colored positive from a set of 
separation negatives, and it uses the oft-discovered fact that when a lantern 
plate is exposed, not all the silver halide grains in that emulsion are used up in 
the formation of the latent or developed image. The plate, after exposure and 
development, is not fixed out, but is thoroughly washed, then bleached with 
ferricyanide. This converts the image into silver ferrocyanide. A subsequent 
treatment with ferric iron, converts it into a blue green. All this is done in the 
dark. The silver halide grains that remain are now exposed a second time, this 
time through another separation negative. The new latent image is then de- 
veloped, and toned to a color complementary to the first. 

108 



KODACHROME AND KODACOLOR 109 

It is very doubtful whether a procedure such as this can be applied in a 
practical manner. The treatment of the plate with oxidizing agents after the 
first image had been developed, no matter how mild, is not without effect upon 
the sensitivity of the remaining unexposed and undeveloped silver halide 
grains. These also are of a speed that is materially different from the grains 
that have been used to form, the first image, since the faster grains will nat- 
urally be used up in the first exposure. The developer oxidation products also 
affect the speed of the unexposed grains, so that all in all the true speed of the 
plate after the first image has been toned a blue, is a thoroughly unknown 
quantity, and a quantity that would not be the same for two successive plates. 
Therefore, while it is possible to obtain some semblance of color by this pro- 
cedure, it would be practically a miracle if accurate color rendition, or any- 
thing even approaching accuracy, could be obtained. However, it is from this 
humble beginning that Kodachrome can be traced. 

Mannes and Godowsky quickly forgot this scheme when they hit upon the 
idea of multi-layered emulsions. The connection is not hard to see, for instead 
of finishing one image before starting the next, they felt that a better control 
would be obtained if the two images were done together. The only practical 
way to insure this, and to be able to color the individual images separately, was 
to segregate them. Then what is simpler than putting two layers one on top 
of the other, and printing an image in each of the two layers? Under these 
conditions it would be simplicity itself to develop the two images simultane- 
ously, and in that way have absolute knowledge that the emulsion speeds in 
the two were completely independent of each other. The conversion of each 
image into a different color, could then be done at leisure under laboratory 
control. This type of emulsion is disclosed in their United States patent 
1516824, which was applied for on February 20, 1923 and granted on Novem- 
ber 25, 1924. Curiously enough, the second patent went through the patent 
office in a much quicker time than the first, so much quicker that it was 
issue4 exactly six months before the other. 

In the second disclosure, a red-sensitive emulsion is coated upon a base such 
as glass or celluloid. Immediately on top of this, is coated an orthochromatic 
emulsion, which has been dyed yellow. The dye prevents the blue light from 
reaching the lower emulsion when the exposure is made through the emulsion 
side of the pack. The blue and the green densities would register upon the top 
layer, while the red densities would register upon the lower one. After ex- 
posure, development, fixation and washing, the pack is treated with ferricya- 
nide. This converts the two images into silver ferrocyanide, which can be 
very easily redeveloped. Next comes the novel feature of the patent. The 
development of the ferrocyanide images was to be accomplished by means of 
a special solution, one whose diffusion into the gelatin could be controlled at 
will, so that only the silver salt in the top emulsion layer becomes reduced. 
No formulas or other description as to how this was to be achieved is given. 
At the conclusion of this treatment, there is present a silver image in the 



no HISTORY OF COLOR PHOTOGRAPHY 

upper layer and a silver ferrocyanide image in the lower. This last could then 
be toned blue with an iron solution (ferric ions must be used, not ferrous as is 
indicated in the specifications), and the upper one could be dye- toned red by 
Traube's silver-iodide mordant process. Of course it would have been just as 
easy to have compounded a ferricyanide bath that had the same properties 
as the second developer, and in that way convert only the upper image into 
silver ferrocyanide. Then a similar situation would result — one image would 
exist in the form of metallic silver and the other in the form of a silver salt. 
Each could be treated by an independent procedure to convert it to color. 

Several years later, a slight variation was disclosed (U.S.P. 1659148; Eng. 
P. 245198). It was realized that because of the presence of the yellow dye in 
the top emulsion layer, that layer itself becomes stratified, for the blue densities 
would be completely restricted to the upper half of this layer, while the green 
densities would exist throughout the entire depth. The red densities would 
still be found in the bottom layer only. The preferred treatment in this case 
was to expose, develop, fix, wash, and treat with ferricyanide. The images in 
the upper emulsion layer are then developed. The red densities, still in the 
form of ferrocyanide, are next converted into an iron blue by the action of 
ferric salts, and the silver images in the upper emulsion are then treated with 
a mordant solution and dyed with a basic magenta dye. Up to this point the 
procedure is pretty much what it was in the original. But if the magenta dye 
image be now subjected to the action of a basic yellow dye compounded so 
that it will not penetrate into the gelatin layer, the upper portions of the 
magenta image will be converted into a yellow. But it is only the upper por- 
tions of this image that correspond to the blue filter densities, hence true 
renditions of these are obtained. Of course complete accuracy is lacking, due 
to the fact that this image contains a considerable measure of green filter 
densities as well as blue. 

Another advance was made in the next disclosure by these men (U.S.P. 
1954452), which was issued in 1934. They had evidently given up the idea of 
obtaining color by toning methods and had turned to color-development, to 
which they subsequently clung with such a tenacity that monopacks and 
color-development appeared to be synonymous. The fundamental principle of 
color-development is that the oxidized developing ions unite with some other 
substance present in the system, to form an insoluble dye. Two possible pro- 
cedures were open. In one case, the emulsions were prepared with the coupling 
substances added. Thus an orthochromatic emulsion was prepared, and to it 
was added a yellow dye and tri chlor naphthol. A red-sensitive emulsion had 
added to it para-nitro-benzyl-cyanide and ethyl-aceto-acetate. The red emul- 
sion was coated upon the subbed base, and the orthochromatic emulsion was 
coated on top of it. After exposure, the pack was developed with a paraphenyl- 
enediamine developer that contained no sodium sulphite. The tri-chlor 
naphthol united with the oxidized diamine to form a cyan dye which was de- 



KODACHROME AND KODACOLOR in 

posited in situ with the silver. The nitro-benzyl-cyanide and the ethyl-aceto- 
acetate also formed dyes with the oxidized developer, this time a mixture of a 
magenta and a yellow dye, so that the net effect was to deposit, beside the 
silver image in the bottom layer, a red dye image. Of course it was necessary 
that the ingredients that were added to the emulsion to effect this dye forma- 
tion, be completely insoluble in water, and very little soluble in the alkalinity 
required for development, otherwise the materials would not stay put in their 
respective layers, and diffusion of the dye, with subsequent loss of definition 
would result. Ansco and the I.G. developed an entire industry in an effort 
to effect this insolubility, but this will be discussed at a later time. The silver 
images were removed by Farmer's reducer, leaving pure dye images. 

In the other procedure, the red-sensitive emulsion was dyed yellow, besides 
having the coupling agents added, and the yellow dye was removed from the 
orthochromatic emulsion. The dye used was such that it dyed the individual 
grain so that any light that struck that grain would first have to be filtered 
through the yellow dye. The grains of the two emulsions were then mixed in 
the proportion required to yield a balanced result. The combined emulsion 
was then coated upon a subbed base. After exposure, development was 
achieved by means of a sulphite free paraphenylenediamine developer, and the 
same result happened. The individual grain had associated with it the sub- 
stance that would give the correct color by coupling with the oxidized de- 
veloper. Other variations of this technique were disclosed in later patents 
(U.S.P. 1980941 and 1997493; Eng. P. 376794, 376795, 376838 and 451699). 

Up to this time, the interest seemed to be centered mainly upon the funda- 
mentals of monopack making and processing, with but little regard to the appli- 
cation to a three-color reproduction scheme. This was made apparent for the 
first time in United States patents 1969469 and 2059884, and English patents 
427472, and 427516. The negative material was a two-layered monopack, 
such as the ones described above, but without having color formers present in 
the emulsion. It was proposed to use this in conjunction with a light-splitting 
camera which would yield two images in adjacent frames of the film. With 
the proper choice of filters, it became possible to register three separations 
upon the two frames. If the exposure on one frame was made through a blue 
filter, then the blue densities would be registered upon the upper layer of the 
pack, since that layer was sensitive to both the blue. and the green. Alternate 
frames were exposed through a blue filter. If the other exposure was made 
through a yellow filter, then in the upper layer of the emulsion there would be 
registered the green densities, while directly underneath it in the lower layer, 
would be registered the red densities. The images in the lower emulsion 
layer were then processed to yield a magenta dye, while those in the upper layer 
were made to yield a cyan dye. 

The monopack was made so that the individual layers were extremely thin. 
Hence the total thickness of the pack was no more than that of an ordinary 



ii2 HISTORY OF COLOR PHOTOGRAPHY 

negative material. This gave a considerable improvement to the definition 
and overall speed of the bottom layer. After exposure, the pack was developed 
in a solution which contained: 



Amidol 


5 parts 


Sodium sulphite 


10 parts 


Sodium bisulphite 


5 parts 


Potassium bromide 


1-4 parts 


Water to 


1000 parts 



The development of the image in a developer of this type appears to start at 
the bottom, hence these are known as depth developers. They are indispensable 
with monopacks, where several layers have to be developed simultaneously 
and at equal rates. After a thorough wash, the pack was fixed, washed again, 
and then treated with the following bleach solution: 

Potassium ferricyanide 10 parts 

Ammonia 28% 10 parts 

Water to 1000 parts 

This converts the three images into easily reducible silver ferrocyanide. The 
next step, following a thorough wash, changed the silver salt to silver, and at 
the same time deposited a magenta image in situ with it. 
This was done by treatment with a solution whose composition was 

Sodium sulphite 0.5 part 

Sodium carbonate 10.0 parts 

Diethyl-paraphenylenediamine hydrochloride 0.5 part 

Water to 1000 parts 

To every 100 cc of this add 0.5 gram of brom-thio-indoxyl, 
dissolved in a little acetone. 

At this point the film contains images in the two layers that are a mixture of 
silver and magenta dye. 

It is desired to treat the film in such a manner that the silver in the upper 
emulsion layer becomes converted into a silver salt, and the dye becomes 
simultaneously destroyed. This last is very easy to accomplish, since the dyes 
formed by coupling developers are all extremely unstable to acids, readily 
decomposing. It is necessary merely to make sufficiently acid the bleach 
solution that will be used to convert the silver into silver halide. The problem 
is how to confine the action of the bleach and the acid to the one layer only. 
This control of the penetration of the bleach is evidently a very important 
item in the processing of monopacks, for the solution of this problem forms the 
subject matter of a rather extensive series of patents (U.S.P. 2019718 and 
2059887; Eng. P. 427518, 454498, 454499 and 454622). It is achieved very 
easily by the inclusion of loading agents such as sodium sulphate, sugar, 
glycerin, alcohol, etc., in the solution. The action of the solution must be 
stopped by means of a solution whose penetration also can be controlled, if 



KODACHROME AND KODACOLOR 113 

that stoppage is to be attained by the use of substances which have a deleteri- 
ous effect upon the dyes. Since the penetration of the solution is to be limited, 
it is desirable that the time of treatment be held to a minimum, so that the 
concentration of reactive agents must be sufficiently high to effect reaction 
immediately. Two such solutions are given. 



A, 


, Chromic acid 10% 


10 parts 




Hydrobromic acid 41 % 


3 parts 




Potassium bromide 


2 parts 




Water 


90 parts 




Alcohol 


300 parts 


B. 


Glycerin 


500 parts 




Isopropyl alcohol 


1000 parts 




Water 


75 parts 




Quinone 


5 parts 




Hydrochloric acid 35% 


20 parts 



These solutions require approximately four minutes to penetrate through the 
upper layer. At the end of this time, the further action is stopped instantane- 
ously by means of an alkaline short stop, such as 



Alcohol 


400 parts 


Sodium sulphite 


20 parts 


Ammonia 28% 


10 parts 


Water to 


1000 parts 



The reformed silver bromide in the upper emulsion layer requires an ex- 
posure to light to make it developable. It is finally developed in a solution 
composed of 100 parts of A and 15 parts of B. 

A. Sodium carbonate 10 parts 
Sodium sulphite 5 parts 
Diethyl-paraphenylenediamine hydrochloride 2 parts 
Water to 1000 parts 

B. 1:3:4 tri-chlor-naphthol 1 part 
Alcohol 150 parts 

With this bath there is deposited a cyan dye image in situ with, and in direct 
proportion to the silver image. The film now contains a magenta dye plus 
silver image in the bottom layer, and a cyan dye plus silver image in the upper 
one. A treatment with Farmer's reducer (which is made by adding hypo to a 
ferricyanide solution), kept alkaline to prevent any effect upon the dyes, re- 
moves the silver from both layers, leaving pure dye images. 

It is to be recalled that alternate frames in this film received blue and 
minus-blue exposures. If frame number one was exposed through a blue filter, 
then frame number two was exposed through a yellow one, and all the odd- 
numbered frames would be identical with the first, and all the even-numbered 



ii4 HISTORY OF COLOR PHOTOGRAPHY 

ones identical with the second. The odd-numbered frames would contain a 
single image, colored cyan, and situated in the upper layer of the film. All the 
even-numbered frames would contain two images — one cyan-colored situated 
in the upper emulsion layer representing the green filter densities; and the 
other magenta-colored, situated in the lower emulsion layer, and representing 
the red densities. Therefore the blue and the green densities lie in the upper 
emulsion layer, and are both colored cyan. They can be printed with red or 
infrared light, which would be modulated by the modulations in the cyan dye 
image. Since these images lie in alternate frames, their separation is very 
easily achieved by printing only alternate' frames upon the same piece of film. 
The red densities can be separated from the others by using a minus-magenta 
light for the printing, that is, by printing with green light. 

It may be wondered why the first development was not made to yield a color 
image directly. This would have saved two processing and two wash opera- 
tions, four steps in all, namely the initial development and the conversion of 
the silver image into developable silver ferrocyanide. The reason for this extra 
labor is that were the development made directly in a color developer, a con- 
siderable loss in emulsion speed and in contrast would have resulted. The 
exact reasons for this we will discuss in a later chapter, when the fundamental 
principles of color development are taken up. Since the monopack was being 
used in a camera, practical requirements demanded that all of the film speed 
be utilized. There is already a sharp loss due to diffusion of the red densities 
by their passage through the front emulsion layer. The extra amount of light 
that is required to yield a sufficiently strong image could be given to the silver 
ferrocyanide or silver bromide formed by the bleach. At this stage, emulsion 
speed is no longer a criterion, for it is possible to use even a carbon arc, if 
necessary, to obtain sufficient exposure. 

The colored negative, prepared as above, was to be printed upon a specially 
coated three-layered monopack, whose structure and processing were disclosed 
in United States patent 2010459, an( i m English patents 427472, 427516, 
427517, 427519, 427520, 440032, 440089, and 441325. On one side of the 
celluloid base there was coated a silver chloride emulsion that was sensitized 
to the infrared rays. This was made waterproof by a coating of varnish, 
made by dissolving 150 grams of benzoyl cellulose (Eng. P. 327714, 339902 
and 356308), in a mixture of solvents whose composition was 

Benzene 1550 parts 

Toluene 100 parts 

Xylene 400 parts 

The varnish could also contain a black pigment, in which case it would act as 
a non-halation coating. But this last was not essential, and would present 
certain complications in the printing. On the other side of the base was coated 
a red-sensitive emulsion, a gelatin layer, and finally a green-sensitive emulsion. 
When the even frames in the colored negative were printed upon this with 



KODACHROME AND KODACOLOR 115 

yellow light, the green component of the yellow would be modulated by the 
magenta image in the negative, and this would register in the green-sensitive 
emulsion of the positive. In a similar manner the red component of the yellow 
printing light would be modulated by the cyan image in the negative and would 
register in the red-sensitive layer in the positive. The blue densities lying in 
the odd numbered frames of the negative are printed with infrared light upon 
the silver chloride emulsion layer. Thus a natural disposition of the three 
primary densities among the three layers in the positive material, is easily 
obtained. 

After the exposures have made been, the film is processed to yield a magenta 
dye image together with the silver. This will take place only in the two un- 
protected emulsion layers in which the red and the green separations have been 
printed, for the other layer is protected from the action of aqueous solutions 
by the varnish. The film is then fixed in hypo to remove the unused, silver 
bromide, washed, and subjected to the action of a bleach whose penetration 
is limited to the upper emulsion layer only. The composition of the magenta 
color-developer and the bleach has been disclosed in a previous specification, 
so they need not be discussed at this time. We have present now a magenta- 
plus-silver image in one emulsion layer; a silver bromide image that requires 
an exposure to light before it can be developed, in the layer immediately 
above it; and a latent image in a silver chloride emulsion layer situated on the 
other side of the emulsion. All the reactions have been conducted in the dark 
up to this time. The next step is to remove the varnish layer from the chloride 
emulsion by means of an organic solvent. This is then developed in a yellow 
color-developer whose composition is 

Diethyl-paraphenylenediamine hydrochloride 10 parts 

Sodium sulphite 5 parts 

Sodium carbonate 20 parts 

Potassium bromide, molar solution 2 parts 

Water to 1000 parts 

To every 100 parts of this solution add 0.1 part of benzoyl- 
acetone, dissolved in a little acetone or alcohol. 

After a thorough wash, the unreacted silver chloride is removed by treatment 
with a 4 per cent solution of ammonia. This will act as a fixing agent for silver 
chloride, but it will have no effect whatsoever upon the silver bromide. Now 
the real reason for the use of this mixed type of emulsion is made apparent. 
Some years later, in 1939, K. Schinzel, the same person who started monopack 
on its long history, took advantage of the differential properties of silver 
chloride and silver bromide, in a large number of disclosures all dealing with 

Imonopacks and monopack processing. 
Now only the silver bromide positive image remains undeveloped. Before 
this can be developed, it must be exposed to light. Development is carried out 



n6 HISTORY OF COLOR PHOTOGRAPHY 

previous disclosure. The silver present in each layer is removed by treatment 
with Farmer's reducer, leaving three dye images. The yellow* image was pre- 
pared by printing through the cyan image in the negative, which represented 
the blue filter densities. Therefore a correct rendition is obtained. The 
magenta image was prepared by the red component of the yellow-light ex- 
posure. This was modulated by the cyan image in the negative. That image 
exists in the upper emulsion layer of the negative, and is the image of the 
exposure that was made through a yellow filter. It, therefore, must represent 
the green densities of the original, hence must be toned a magenta in the 
positive reproduction. A similar analysis will show that the cyan-colored 
image in the copy arose from the magenta image in the negative, and this is 
due to the red densities in the original. 

The disclosures up to this point do not correspond to the use of a film which 
has emulsion coated only on one side, and which is processed by reversal. But it 
is not a great step from the two-layered negative to a monopack material that is 
capable of giving three dye images in three individual layers. This is disclosed in 
United States patent 2113329, and English patents 444198, 447092, and 455128. 
A transparent base of celluloid has an anti-halation layer coated on one side, 
and on the other, three emulsion layers separated by two gelatin layers. The 
bottom layer is a red-sensitive emulsion, made by the inclusion in it of naphtho- 
carbocyanines. Above it is a layer of clear gelatin which serves to differentiate 
sharply the red-sensitive from the green-sensitive layer, and also to enable the 
green emulsion to be coated easily upon the other. It is very desirable that 
no intermixture of sensitized grains take place, and no diffusion of excess 
sensitizing dye from one zone to the next. The middle emulsion layer is 
sensitized by means of erythrosin. Above this is a layer of yellow-dyed gelatin, 
and this is finally topped with a color-blind emulsion. This last could also be, 
and preferably so, a very fine-grained emulsion that has had its blue sensitivity 
increased by the use of blue sensitizing dyes. This serves two purposes. The 
main one is that such an emulsion would allow a much better transmission 
through it of the minus-blue light, since a fine-grain emulsion would not diffuse 
nor scatter that light in anywhere near the amount that a normal fast emulsion 
would. Another purpose would be to cut down as much as possible the ratio 
of the green sensitivity to the blue. All color-blind emulsions that have suffi- 
cient speed to be used for negative purposes, contain silver iodide, and this salt 
acts as an optical sensitizer for the emulsion, bringing the sensitivity up to 
approximately 520 mju. The amount of the green sensitivity is not very great, 
but still enough to render certain shades of blue-green much lighter than 
proper. If the optical sensitizer has an absorption that falls short of 500 m/x, 
then the ratio of the green sensitivity beyond 500 to the blue sensitivity up to 
500, is reduced considerably. No blue light can penetrate beyond the yellow 
filter layer beneath the top emulsion layer. During the processing, the gelatin 
emulsion absorbs considerable water, probably twenty times its own weight. 
This results in a ten-time increase in the thickness of the layer. Therefore, 



KODACHROME AND KODACOLOR 117 

the presence of gelatin layers between the emulsions serves the third purpose 
of making much less critical the time of treatment with those solutions whose 
penetration into the emulsion layers must be controlled. 

The initial exposure of such a film results in a clean-cut separation of the 
three primaries among the three layers of the pack. The blue densities are 
registered in the top layer only; the green in the middle; and the red in the 
bottom. These latent images are developed in a typical reversal developer, one 
that contains a silver-halide solvent such as hypo or thiocyanate. These 
developers will be discussed later when the screen plate processes are taken up. 
Such a developer is the following: 



Metol 


6 parts 


Sodium sulphite 


75 parts 


Hytfroquinone 


10 parts 


Sodium carbonate 


30 parts 


Potassium thiocyanate 


1.75 parts 


Potassium bromide 


2\ parts 


Formalin 40% 


2§ parts 


Water to 


1000 parts 



After development is complete, the film is washed thoroughly, then treated 
with acid permanganate to remove the developed silver image. This solution 
is made as follows: 

Potassium permanganate 4% 5 parts 

Sulphuric acid 20% 5 parts 

Water to 100 parts 

Another wash removes excess permanganate, then a bath in 2 per cent sodium 
bisulphite removes the permanganate reduction products, and restores the 
light-sensitivity of the remaining silver-halide salts. All of the treatments, 
except the development, require approximately four minutes time at 70 F. 
But the temperature of the bleach bath should be maintained at 65 F to pre- 
vent any too great softening of the gelatin. After an exposure to light whose 
intensity is sufficient to make all the remaining grains developable, the film is 
developed in a cyan color-developer such as: 

A. Diethyl-paraphenylenediamine hydrochloride 3 parts 
Sodium sulphite 5 parts 
Sodium carbonate 50 parts 
Potassium thiocyanate \ part 
Water to 1000 parts 

B. Meta-hydroxy-diphenyl 2% parts 
Methyl alcohol 100 parts 

Add B to A 

In this solution there was deposited a cyan dye together with the silver, in 
each of the three emulsion layers. The film was then washed, fixed, washed, 
and dried in preparation for the second stage. 



n8 HISTORY OF COLOR PHOTOGRAPHY 

This consisted of the conversion of the silver images in the two upper layers 
into silver chloride, with the simultaneous destruction of the cyan dye image. 
It was accomplished by treating the film with the following solution for four 
minutes at 72 F. 

Glycerin 500 parts 

Isopropyl alcohol 1000 parts 

Water 75 parts 

Quinone 5 parts 

Concentrated hydrochloric acid 20 parts 

The action of this bleach was stopped by immersion in a stop bath whose 
composition was: 

Sodium bicarbonate 15 parts 

Isopropyl alcohol 1000 parts 

Glycerin 1000 parts 

Water 1000 parts 

The time of immersion was ij minutes. After a thorough wash, the film was 
developed in a magenta color-developer. 

A, Diethyl-para-toluylene-diamine hydrochloride 1 part 
Sodium sulphite 10 parts 
Sodium carbonate 30 parts 
Potassium thiocyanate J part 
Water to 1000 parts 

B. Para-nitro-phenyl-acetonitrile J part 
Acetone 20 parts 
Alcohol 100 parts 

Add B to A 

After a wash, the film is dried in preparation for the third stage. 

The film now contains a cyan-plus-silver image in the bottom layer, and 
magenta-plus-silver images in the other two. It is desired to convert the image 
in the top layer into a yellow. This is the function of the third stage. The 
magenta dye in this layer is destroyed, and the silver simultaneously con- 
verted into silver chloride by treatment with a bleach compounded like the 
one above. The action of this bleach is stopped in a stop bath, also like the one 
disclosed in the second processing stage. After a thorough wash the film is 
given its last development in a yellow color-developer. 

A . Di-methyl-paraphenylenediamine sulphate 1 part 
Sodium sulphite 2 parts 
Sodium carbonate 30 parts 
Water to 1000 parts 

B. 4 nitro-aceto-acetanilide 2 J parts 
Alcohol 100 parts 

Add B to A 



KODACHROME AND KODACOLOR 119 

The silver can be removed by means of Farmer's reducer, leaving pure dye 
images. These are a cyan in the bottom, a magenta in the middle, and a 
yellow in the top layer. 

The box in which the 35 mm film is packaged carries on its cover a list of pat- 
ent numbers, which presumedly disclose the structure and the processing of the 
material. The structure is supposedly disclosed in United States patent reissue 
18680; for the other two numbers, 1638577 and 2019672, deal with the struc- 
ture of a cartridge to hold the film. Reissue 18680 we have already discussed 
in the preceding chapter. It was issued to Dr. Troland, and assigned to Techni- 
color. It deals in general with a two-layered monopack film, with or without 
a filter layer between the emulsions. Nowhere in the specifications is there 
discussed the use of the material to yield a reversed or direct three-color trans- 
parency. It is certainly not an easy matter to trace any connection between 
the structure of the Kodachrome film and the structure of a Troland mono- 
pack. It is only in the reading of a few of the claims, worded in a very general 
manner, that any connection may exist. One such general claim deals with the 
existence of color-complemental latent images, in different strata of an emul- 
sion. Another claim deals with the existence of color-complemental silver 
images in different strata of an emulsion. But a claim has validity only in so 
far as it is fully described in the specifications, and these mention only images 
formed in a two-emulsion-layered monopack. Even if the more general view 
be taken, there is always the prior art to fall back on. In a Schinzel type of 
monopack (of last chapter), after the exposure has been made, it is impossible 
to have anything but three color-complemental latent images existing in 
different strata of an emulsion. And after development, it is impossible to have 
anything else except three color-complemental silver images present in differ- 
ent strata of the emulsion. 

If this is not sufficient, then there is the patent (Eng. P. 15055/12) issued to 
R. Fischer, twenty years before Trolajid. Among other things, there is de- 
scribed a monopack in which three emulsions are coated one on top of the other. 
Between the top blue-sensitive emulsion, and the middle green-sensitive one, 
there is placed a gelatin layer dyed yellow. A colorless layer of gelatin is 
placed between the green- and the red-sensitive layers. We will discuss this 
patent in greater detail when Ansco Color is taken up. Here it need only 
be mentioned that a pack substantially like Kodachrome was disclosed in 191 2, 
hence such could no longer be deemed a basic invention in 192 1. The Schinzel 
disclosure was cited against Troland, but he was allowed the claim when he 
answered that Schinzel never put his disclosures into practice. The Fischer 
patent was not cited, hence no answer was required. This, of course, merely 
means that the Troland disclosures must still be tested from a legal point of 
view, in light of the more substantial Fischer disclosures. 

The 35 mm film is evidently processed in accordance with United States pat- 
ents, 1460703, 1900870, 1939231, 1954346, 1980941, and 1997493. Of these, only 
the last two were issued to Mannes and Godowsky. The first of these, 1980941, 



120 HISTORY OF COLOR PHOTOGRAPHY 

deals with an emulsion which contains a mixture of sensitized grains. Thus a 
blue-sensitive emulsion is mixed with a red-sensitive yellow-dyed emulsion, and 
the mixture coated upon a celluloid base. After exposure in the camera, the 
film is developed in an ordinary developer, and then washed. This procedure 
does not destroy the sensitivities of the dyed grains to red light, so that an ex- 
posure of the film at this stage to red light will produce a positive latent image 
only upon the red-sensitive grains. The exposure should be sufficient to make 
all the remaining red-sensitive grains developable. This last can be accom- 
plished with a cyan color-developer. It is followed by a white- or blue-light 
exposure to make the color-blind grains developable, and these are developed 
to a red color. The silver images, both negative and positive, can be removed 
by Farmer's reducer. The procedure can also be applied to three color. The 
other Mannes and Godowsky disclosure (U.S.P. 1997493) relates to the proc- 
essing of a two-layered monopack film. The upper layer can be developed using 
a surface developer, and the silver image here can be toned blue. Then the 
bottom latent image is developed and dye-toned to a complementary color. 
Or, both latent images are developed simultaneously, and the upper one toned 
blue with a solution compounded so that its diffusion could be controlled. The 
lower image is then dye-toned to a complementary color. 

United States patent 1460703 deals with reversal in general, and has no 
reference to the treatment of color film. But the technique of etching out a 
silver image and exposing the remaining silver salts, is disclosed. This 
probably is used in Kodachrome processing. United States patent 1900870 was 
issued to M. Seymour. This discloses the use of an emulsion sensitized with a 
dye that withstands the action of acid dichromate. A monopack is coated with 
such an emulsion for one of the layers, say the red-sensitive one. After expo- 
sure, development, and removal of the silver by a bath in acid dichromate, the 
film is bathed in sodium bisulphite, exposed to red light, and color-developed 
with a cyan developer. This uses up the remaining red-sensitive grains. Ex- 
posure to white light followed by development with a complementary color- 
developer, completes the processing. The silver is removed by Farmer's re- 
ducer. Another Seymour patent (U.S. 1939231) included in the processing list, 
deals with reversal. The film after exposure and development in a non- 
staining developer, is re-exposed to light and developed in a blue-black or almost 
neutral gray color-developer. Two such are listed: 



1. Alpha naphthol 


i part 


Acetone 


25 parts 


Para-amino-diethyl-aniline 


Ipart 


Sodium carbonate 


7 J parts 


Water 


250 parts 


2. 1 : 5 di-hydroxy-naphthalene 


§ part 


Acetone 


20 parts 


2 : s di-brom-4-amino-phenol 


i part 


Sodium carbonate 


7| parts 


Water 


250 parts 



KODACHROME AND KODACOLOR 121 

The remaining patent in the list is one dealing with a two-layered monopack, 
in which one layer is sensitized with a stable dye. This patent was issued to 
CapstafT. It discloses a technique that is very much akin to that disclosed by 
Seymour. These patents do not appear to be wholly consistent with each 
other, and they certainly are considerably different from the description of the 
Kodachrome process outlined by E. R. Davies, a member of the British Kodak 
Research Laboratories (Phot. J., Vol. 76 (1936), p. 248). In this paper the 
technique appears to be that disclosed in United States patent 21 133 29. 

Kodachrome cut film was introduced to the industry a year or two after the 
35 mm film. On the box in which the film is packaged, patent reissue 18680 
is noted, and this has already been discussed. The other processing patents 
noted on the package are 1460703, 1939231, 1954346, 1980941, 1997493 (all of 
which appear also on the 35 mm boxes), 1966330, 2019718, 2059887 and 
2113329. One of these (1966330), issued to Burwell, discloses a method of 
processing a film that has single layers of emulsion coated on both sides. This 
is the ordinary double-coated or duplitized positive film, so popular in the old 
two-color days. A primary separation is printed on each side of the film, and 
it is processed to yield silver-plus-cyan dye images. Then one side only is 
treated, after fixation and wash, with : 



Potassium f erricyanide 


2 parts 


Potassium bromide 


2 parts 


Sulphuric acid 


J part 


Water to 


100 parts 



This destroys the cyan dye and converts the silver into silver bromide. 

Redevelopment with a red developer follows. The silver is removed by 
Farmer's reducer, leaving two pure dye images. The developers used have the 
following compositions: 



A . Cyan Developer 




3 : 5 di-brom-ortho-cresol 


2 parts 


Ethyl alcohol 


20 parts 


Sodium carbonate 


7 s parts 


Potassium bromide 


0.3 part 


Para-amino-diethyl-aniline hydrochloride 


2 parts 


Water to 


250 parts 


B. Red Developer 




Cyan-aceto-phenone 


1 part 


Ethyl alcohol 


20 parts 


Ethyl-alpha-chlor aceto-acetate 


1 part 


Sodium carbonate 


7! parts 


Para-amino-diethyl-aniline hydrochloride 


1 part 


Potassium bromide 


0.3 part 


Water to 


250 parts 



If the ethyl-aceto-acetate be left out, a magenta rather than a red image would 
result. From this disclosure it does not seem that it is an important step in 



122 HISTORY OF COLOR PHOTOGRAPHY 

Kodachrome processing. The last named patent (21 133 29) is apparently the 
true description of what actually takes place in the laboratory after the film 
is received there. The other two disclosures (2019718 and 2059887) deal with 
the all-important steps of controlled penetration of the bleach solutions into 
the film. These have been discussed in some detail above. 

Some light as to the exact processing methods used was given by Dr. C. E. K. 
Mees, vice-president of Eastman Kodak Company, and the director of research 
{Am. Phot. Vol. 36 (1942) Mar. p. 8; Communication No. 832, Kodak Re- 
search Laboratories) in a paper that disclosed just what Eastman was doing 
in the field of color. According to Dr. Mees, Kodachrome was no longer 
processed by the technique developed by Mannes and Godowsky, which 
utilized a controlled diffusion of a bleach bath. The newer method was to 
utilize the residual color sensitivity of the layers after they had been exposed 
and developed, to effect a layerwise separation of the complemental images. 

At this point the film contains three silver negative images imbedded in 
the three layers of unexposed silver bromide. The red and green sensitizers 
have been left unaffected by the development, hence the central layer re- 
mains green-sensitive and the bottom layer remains red-sensitive even after 
the exposed portions of these layers have been completely reduced. An ex- 
posure to red light effects a reversal only in the bottom layer. Upon develop- 
ment with a cyan color-developer of the type noted above, a cyan dye plus 
silver image is formed in the bottom layer only. The exposure and develop- 
ment must be sufficient to effect complete reversal, so that after this treatment 
the bottom layer no longer contains any light-sensitive elements. The next 
step is to expose the top layer with blue light, and color develop this with a 
yellow coupler-developer. The yellow-filter layer existing between the blue- 
sensitive and the green-sensitive emulsions must be insensitive to the developer 
baths, for it is essential that this layer be present during the time the top 
layer is completely fogged by the blue light. Otherwise a spill of blue exposure 
into the green layer would take place, and this would destroy the quality 
of the final color image. 

After the blue layer has been reversed, there remains only the central layer. 
This contains a silver negative plus a silver halide positive image. It can be 
made developable either by exposure to light, or by treatment with such fogging, 
agents as methylene blue, thiourea, etc. Development is effected by means 
of a magenta coupler-developer. It remains now only to remove the silver 
images by treatment with Farmer's reducer or some other relatively mild 
bleaching agent which would leave the dye images unaffected. 

In the same article, Dr. Mees disclosed another Eastman color process called 
Kodacolor. Like Kodachrome, the name had previously been used to denote 
an entirely different color process, that had been discarded. This will be 
discussed in a later chapter dealing with the lenticular processes. Kodacolor 
is a monopack film which yields a complementary colored negative as the 
direct result of the camera exposure. Strictly speaking, the discussion of 



KODACHROME AND KODACOLOR 123 

Kodacolor should be postponed until the subject of Ansco Color is taken up> 
since it has some properties in common with that product. But we will forego 
consistency for convenience, and continue our discussion at this point. 

In Kodachrome, the couplers were added to the developer solutions. This 
made necessary the formulation of a processing technique which involved a 
stepwise processing of each layer. The film was developed in four separate 
stages. First was the normal development of the original exposure. Next 
there was the development of the cyan image in the bottom layer, this being 
followed in turn by a similar processing for the top and the central layers. 
In Kodacolor a different procedure was adopted. The couplers were made 
part and parcel of the film itself. Each layer had associated with itself a coupler 
which would unite with the oxidized developer to deposit a pigment or insoluble 
dye. Ansco and the I. G. had already disclosed one method to accomplish 
this (cf. Chapter n). The procedure was to convert the coupling agents into 
non-diffusing components by alteration of the structure of the molecule. 
Eastman adopted the procedure of grain isolation. The couplers were dis- 
persed in water-insoluble but water-permeable resins, cellulose esters, ethers, 
or other polymers. The new bodies were then dispersed in the emulsion 
layer. The structure of a Kodacolor layer is quite different from anything 
that had previously been proposed. The individual silver halide grains are 
dispersed in the gelatin and isolated from each other by walls of gelatin. 
Interspersed between them are resin particles. These also are insulated from 
each other and from the silver halide grains, by walls of gelatin. There is, 
therefore, considerable granularity to the layer. 

When an exposed grain is developed, the oxidized developer diffuses away 
from the developed grain. In its travels it will permeate the resin and once 
inside, will combine with the coupler present there. Therefore, a Kodacolor 
image cannot be as sharp as a Kodachrome, since the dye image is removed 
from the area where the light was originally focused. This effect might be 
offset by a more uniform distribution of the dye particles, hence a smoother, 
more homogeneous image might result. 

Kodacolor evidently is based upon the disclosures of M. Martinez. In his 
first disclosure (U.S.P. 2269158; Eng. P. 505834) he proposed the general 
solution whereby the coupler was dissolved in a resin, and the result dispersed 
in the emulsion. In a later patent (U.S.P. 2284877; Eng. P. 543606) he sug- 
gested that the silver halide grain also be dispersed in the resin, so that each 
resin unit contained all the ingredients necessary for image formation. 

The extension of this principle was soon taken up by other members of the 
Eastman staff. Mannes and Godowsky (U.S.P. 2304940; cf. also Eng. P. 
524555) suggested cellulose esters as being suitable for coupler isolation. In 
English patent 524554 it was suggested that agents be added to the resin 
solutions to increase their porosity. The surface could also be treated to re- 
duce light scatter since the index of refraction of the resin would be different 
from that of the gelatin. S. S. Fierke suggested other substitutes for the 






124 HISTORY OF COLOR PHOTOGRAPHY 

resin, notably polyvinyl acetate, polyvinyl acetal or polystyrene (U.S. P. 
2272191). A diffusing coupler could be converted into one that was diffusion- 
fast by adding a water-soluble resin such as polyvinyl alcohol or phthalate. 
This suggestion was made by Peterson (U.S.P. 2289803). Methods of dis- 
persing the resin in gelatin or silver halide gelatin emulsions, were disclosed by 
Bennett and Fierke (U.S.P. 2311020) and E. £. Knott (Eng. P. 540366, 
S40367, 540368, 544I35)- 

The inverse procedure, the dissolution of a color former in gelatin, then dis- 
persion of the gel into a cellulose ester silver halide emulsion, was disclosed by 
Leermakers and Fierke (U.S.P. 2279406; 2318788; cf. also English patent 
524154). Such a procedure might be suitable for print or duplication purposes, 
since it is quite improbable that sufficient sensitivity could be obtained with 
other than gelatin emulsions. 

The processing of Kodacolor involves merely a development with a solution 
containing practically no sulphite, and some derivative of paraphenylene- 
diamine as the reducing agent. Color is deposited as the silver is reduced. 
This is followed by a wash then by a bleach bath compounded so as not to 
harm the dye images. Fixation in hypo, and a wash complete the processing. 



CHAPTER 11 
ANSCO COLOR AND AGFACOLOR 



Wh: 



'HILE the Eastman Kodak staff has developed to a high degree of pre- 
cision the controlled diffusion methods for processing monopack film, work has 
been going on apace on another procedure for obtaining a like result. Curi- 
ously enough, this method is the one originally disclosed by Fischer and 
Siegrist (Eng. P. 15055/12; U.S.P. 1055155; Ger. P. 257160). In accordance 
with these disclosures, three emulsions were prepared that were sensitized to 
the blue, blue and green, and blue and red respectively. It was not possible 
to prepare a silver-halide emulsion whose sensitivity to the blue was repressed 
to a point where it would no longer matter. The three emulsions contained, 
besides the sensitizing dyes, the proper coupling agents, so that in the blue- 
sensitive emulsion there would be included substances which would yield a 
yellow dye by condensation with the oxidized developer; in the green-sensitive 
emulsion there would be a magenta, and in the red-sensitive emulsion a cyan, 
coupling-agent. The three emulsions were to be coated one on top of the 
other, with a layer of gelatin, dyed yellow, between the blue-sensitive and the 
other two layers, and a layer of clear gelatin between the green- and the red- 
sensitive layers. In this form, the disclosure might seem a prior conception of 
the monopack film as made in accordance with the description in the Mannes 
and Godowsky United States patent 21 133 29. The only difference between 
the two is that the earlier patent disclosed the inclusion of the dye couplers in 
the monopack, while the other preferred to put the couplers into the developer 
solutions. However, it is to be recalled that in one of the earlier Mannes and 
Godowsky efforts (U.S.P. 1954452) it is also suggested that the couplers be 
added to the emulsion directly. The green-sensitive emulsion has added to it 
para-nitro-benzyl-cyanide and ethyl-aceto-acetate, and the red-sensitive emul- 
sion has included in it tri-chlor-alpha-naphthol. After exposure, the pack is 
developed in a developer containing para-amino-di-ethyl-aniline or some 
equivalent substance. The ingredients in the green-sensitive layer combine 
with the oxidized developer to yield magenta and yellow dyes, so that the com- 
bined result would be red, while the ingredient present in the red layer yields a 
cyan dye. Treatment with Farmer's reducer removes the silver, leaving pure 
dye images. Thus it would appear that not only was the basic idea of the in- 
clusion of the couplers in the emulsion layers previously disclosed by Fischer 
and Siegrist almost twenty-five years prior to Mannes and Godowsky, but that 
the type of ingredients used as couplers was also identical with that disclosed 
in the very early patents. 

125 



126 HISTORY OF COLOR PHOTOGRAPHY 

No specific directions were given as to how these substances were to be put 
into the emulsion layers so that a stable result would be obtained. Fischer 
and Siegrist included the gelatin layers between the emulsions not only to 
serve as light filters, but also to completely segregate one layer from the next, 
so that the coupling agents in one zone would not interfere with those in an- 
other zone. This has evidently failed to satisfy the requirements for complete 
segregation, for the pack described by the early inventors never achieved com- 
mercial success. It is probably this slight difficulty that mainly directed the 
work of the Kodachrome group into controlled diffusion processes. But in some 
later patents, Messrs. Mannes and Godowsky return to this field, and in United 
States patent 1996928 they take advantage of the fact that when silver bro- 
mide is prepared in the presence of substances like alpha naphthol and its de- 
rivatives, or para-nitro-benzyl-cyanide and its substitution products, these will 
become adsorbed to the silver halide grain. Thus it becomes possible to pre- 
pare an emulsion in which each grain contains associated with itself the coupling 
agents that are required to form the proper positive color, upon condensation 
with oxidized developer. This adsorption, they claim, does not interfere 
with the sensitization of the grain with sensitizing dyes. A red-sensitive 
emulsion can therefore be prepared with alpha-naphthol derivatives adsorbed 
upon the grains, and a green-sensitive emulsion with adsorbed benzyl-cyanide 
derivatives. These can then either be coated in separate layers one on top of 
the other, or they can be tanned to isolate each individual grain, then mixed 
and coated as a single layer. In the first case it will suffice to include between 
the two layers a yellow-dyed filter layer. But in the second case, each grain 
of the red-sensitive emulsion must be dyed with a yellow filter dye, which will 
become mordanted to the gelatin surrounding the grain. Since this has been 
tanned, it is not a very difficult thing to accomplish. An idea such as this, but 
not to be used in conjunction with color formers, was advanced as far back as 
191 1, by Sforza and a little later by Luther. In a later disclosure (Eng. P. 
503845) it is proposed to use color formers of the pyrazolone group. These 
have a labile hydrogen in the molecule, and can form silver salts that are very 
insoluble. The purified silver salt is dispersed in gelatin and then treated with 
potassium bromide. This converts the silver pyrazolone salt into silver 
bromide, and leaves the pyrazolone adsorbed upon the halide grain. 

To prepare such an emulsion, the following procedure was disclosed: Two 
solutions were prepared, one containing a solution of pyrazolone in ammonia, 
and the other a solution of silver nitrate in ammonia. 

A, i-phenyl-3-methyl-5-pyrazolone 17.41 parts 
Ammonia 28% 30 parts 

B. Silver nitrate 8.495 P ar *s 
Water 170 parts 
Strong ammonia sufficient to redissolve the 

precipitate 



ANSCO COLOR AND AGFACOLOR 127 

Solution A is heated on a water bath to drive off the ammonia until precipita- 
tion of the pyrazolone salt starts, at which point 150 cc of water is added. 
Solutions A and B are then mixed at room temperature. The precipitate is 
allowed to stand overnight, the liquid decanted off, and the precipitate washed 
by decantation and centrifuging. It is finally dried in a desiccator at 55 C. 
The emulsion is made from this precipitate. Two solutions are again prepared. 

A, Silver pyrazolone salt 1.8 parts 
Water 10 parts 
Gelatin 15% 1 part 

The salt is broken up, and suspended in the gelatin at 45 C. 

B, Potassium bromide 0.5 part 
Water 3 parts 

B is added slowly to A with constant stirring, over a period of fifteen minutes. 
Then 15 cc of 15 per cent gelatin is added, and the emulsion is stirred. It is 
then set, shredded, washed for two hours, drained overnight, remelted, filtered, 
and coated. The speed of the plate is approximately the same as a lantern 
plate. It can be developed in a para di-ethyl-amino-aniline developer. 

The fact that the later Eastman work deals with emulsions to which couplers 
have been added, would seem to indicate that this procedure has considerable 
merit. It remained for Ansco and the I.G. to develop it to a successful 
conclusion. But it was not until after Kodachrome appeared on the market 
that they proceeded to develop their product, known as Ansco Color and 
Agfacolor Neue. The first English patent, issued in 1936, bears the number 
455556. Here is disclosed a scheme to obtain an even dispersion of the 
couplers in the emulsion layers. This is done with the aid of wetting agents, 
such as the sulphonated hydrocarbons, sulphonated castor oils, etc. A blue- 
sensitive emulsion has added to it, first a wetting agent like sulphonated 
taurine (sold by the I.G. under the name Igepon), then an alkaline solution 
of tere-phthaloyl-bis-acet-anilide. In the green-sensitive layer, with the aid of 
sulphonated castor oil, is dispersed the alkali salt of the symmetrical deriva- 
tive of urea, prepared from para-amino-benzoyl-meta-amino-phenyl-methyl- 
pyrazolone. 

But the true problem in the preparation of such layers is not the inability 
to obtain uniform dispersions of the couplers throughout the emulsion, but to 
prevent the couplers in one emulsion layer from wandering over into a neigh- 
boring layer. By far the greatest efforts have been expended upon the solu- 
tion of this problem. One method by which this can be done (Eng. P. 458400), 
is to attach substantive groups to the molecule, such as the residues of stilbene, 
azoxybenzene, di-phenyl, oxy-naphthoyl amides, benzthiazol, phenols with 
substitutions in the one and three positions, and amino phenols. The coupler 
molecules will under these conditions attach themselves to the gelatin mole- 
cule, and apparently "dye" the gelatin with colorless and transparent "dyes." 
The preparation of such substances is covered in English patents 365531, 



128 HISTORY OF COLOR PHOTOGRAPHY 

403915, 415945, and 4 I 793°- A typical monopack would contain three layers 
coated one above the other with gelatin filter layers between the emulsions, 
each properly sensitized and each containing diffusion-fast coupling agents. 
The red-sensitive emulsion contains 3:5 di-(phenyl-amino)-phenol. The 
green-sensitive layer contains para-(oxy-naphthoyl-amino)-phenyl-3-methyl- 
5-pyrazolone. The blue-sensitive emulsion has added to it tere-phthaloyl-bis- 
acetanilide. One gram of the coupler is dissolved in a little alcohol, then dis- 
persed in 50 cc of 10 per cent gelatin. The dispersion is finally added to 100 cc 
of the sensitive emulsion. 

In another disclosure (Eng. P. 464398), a slight variation is introduced. The 
yellow-image layer contains either tere-phthaloyl-bis-acetanilide, or benzoyl- 
acetanilide. The magenta layer contains i-(di-phenyl)-phenyl-3-carbethoxy- 
5-pyrazolone. The third layer contains di-(2 : 8-di-hydroxy-3 -naphthoic acid)- 
benzidide. After exposure and color development, there is formed a 
silver-plus-dye image in each of the emulsion layers. But the dye image is 
relatively low in concentration, since for every four atoms of silver that are 
formed, but one molecule of dye is deposited. This results in a loss of speed 
and of contrast, after the silver image is removed. To offset this tendency, the 
silver is converted into a silver anti-diazotate salt, by first converting it to 
silver chloride, then treating the chloride with sodium anti-diazotate. In 
this form the diazo derivative cannot unite with the couplers present in the 
emulsion. Therefore it is possible to wash out completely all the diazo com- 
pound from the non-image portions of the film, leaving it only at image por- 
tions in the form of a silver salt. If the pack at this point be treated with a 
weak acid, the anti-diazotate becomes converted into the normal diazonium 
salt, and this immediately condenses with the coupling agents present in the 
film. Thus for every atom of silver that is formed by the development of the 
latent image, there will be formed one molecule of dye. 

The formation of diffusion-fast couplers is tackled again in English patents 
465823 and 488048. This time long-chain hydrocarbon residues are attached 
to the coupler molecule, as for example in the form of an amino substitution. 
This is utilized in the formation of a pack as disclosed in English patent 47 5 191. 
To 1000 parts of emulsion is added one or the other of these solutions: 

To blue-sensitive emulsion. 

Di-(benzoyl-acetic-acid)-benzidide 10 parts 

Alcohol 30 parts 

Sodium hydroxide, 20% aqueous solution 10 parts 

To green-sensitive emulsion. 

i-(para-stearyl-amino-phenyl)-3-methyl-5-pyrazolone 10 parts 

Methyl alcohol 10 parts 

Sodium hydroxide, 20% aqueous solution 10 parts 

To red-sensitive emulsion. 

Di-(i-oxy-2-naphthoyl)-benzidide 10 parts 

Methyl alcohol 10 parts 

Sodium hydroxide, 20% aqueous solution 10 parts 



AN SCO COLOR AND AGFACOLOR 129 

Rapid filter yellow could be added to the color-sensitized emulsions. The pack, 
after exposure, is developed in a para di-ethyl-amino-aniline developer. 

In disclosure 476672 all the three layers of the pack contain the long-chain 
hydrocarbon residue. The yellow layer contains para-capronyl-amino-benzoyl- 
acetamino-salicylic acid. The cyan layer contains 1 -hydroxy- 2-carboxy-5- 
dodecoyl-amino-naphthalene. The magenta layer contains i-(4 / -oxy-3 / -carboxy- 
phenyl)-3-(4 // -dodecoyl-amino)-phenyl-5-pyrazolone. This did not complete 
the possibilities for the formation of diffusion-fast couplers, for they return to 
this problem again and again in a large number of patents. Thus it is proposed 
to use highly polymeric carboxylic acid derivatives (Eng. P. 479838) such as 
polyglucuronic acid, polyvinyl carboxylic acid, etc., as substitutions within 
the coupling molecule. In English patent 483000, it is suggested that sugars 
be used, and that one or more of the free hydroxy groups in the sugar molecule 
be esterified or etherized with coupler residues. It has also been proposed to 
use the degradation products of gelatin, albumen, or other colloids, causing 
these to combine with the coupling agents through a carboxyl or an amine 
group. The addition of a sterol, cholesterol, etc., group to the coupler through 
union with a carboxylic acid or an amino group, is described in English patent 
489093. The presence of several coupler residues in a single molecule is again 
disclosed in English patent 4891 61. Recourse to heavy molecules is resorted to 
in English patent 489164. Here diffusion-fast couplers are made by replacing 
one or both of the hydroxyl groups in malonic acid (HOCO * CH 2 • COOH) 
with groups like NHR, where R is a heavy radical which would make the 
entire compound substantive to gelatin. Examples of such substances are 
C 6 H 5 NH— CO— CH 2 — CO— NHC8H15 and . 

C 6 H 5 — C 6 H 4 — NH— CO— CH 2 — CO— NH— C 6 H 4 — C 6 H 6 . 

The heavy hydrocarbon attached to the nitrogen atom, can also contain a 
carboxyl or sulphonic acid substitution, to give the molecule water solubility 
in the form of an alkali metal salt. Resin residues (Eng. P. 489274), or hydroge- 
nated aromatic residues (Eng. P. 491958) when substituted into the coupler 
molecule, also gave it substantive properties. 

In the preparation of these substances it is desirable to have the compounds 
soluble in the alkalinity required for development with substituted para- 
phenylenediamine compounds. This, at first glance, may seem to be incon- 
sistent with the requirements for diffusion-fastness, with the requirement that 
the molecule stay put in the position where it is lodged. But here good old 
colloidal chemistry comes to the rescue. Diffusion through a gelatin mem- 
brane is dependent to a large extent upon the particle size and molecular weight 
of the diffusing molecule. The larger the molecule and the greater its molecular 
weight, the less likely it is to wander to any great extent, in any given time. 
It is for this reason that solutions of the simple inorganic salts, whose molecular 
weights are fairly low, diffuse with great ease through gelatin, and may be 
washed in or out of a gelatin layer. But when the molecular weight reaches 
well into three figures, the molecule becomes too heavy to respond to Brown- 



130 HISTORY OF COLOR PHOTOGRAPHY 

ian movement, or even to osmotic pressures. For that reason heavy dye 
molecules, once they stain gelatin, are very difficult to remove. This is taken 
advantage of in imbibition processes. There it is desired to have a water- 
soluble dye which has a tendency to stain gelatin, become imbibed by a matrix. 
This is a gelatin relief carried by a celluloid base. The relief, at its deepest 
point, is of the order of o.ooooi inch. After dyeing, it is brought into contact 
with a plain gelatin film whose thickness may be from ten to one hundred 
times as much, and whose physical hardness is much lower than that of the 
matrix. The two are brought into contact so that a layer of water, extremely 
thin, but still present, lies between them. The water dissolves out some dye 
from the matrix, since that dye is water-soluble. The solution then dyes the 
blank, since that is receptive to dyes. If the molecule is very large, it will 
move very sluggishly, once it is in the gelatin film, and hence excellent defini- 
tion becomes possible. If the conditions of the water interlay er are made such 
that the solution of the dye becomes molecular and crystalloidal rather than 
colloidal, then the transfer will have very poor definition. This will take place 
if it is made alkaline. Much more of the dye will transfer, and in a much 
shorter time, but it will not stay put in the gelatin layer, and will continue to 
wander about until it has been made sufficiently dry to stop its traveling. 
Treatment of the dyes with acids like acetic, is also based upon a similar idea. 

We meet the same influence in the conversion of the silver image into a 
silver iodide mordant image by means of a bleach of the Miller type. There 
the iodizing solution is compounded to contain at least 7 per cent of potassium 
iodide. In so strong an iodide solution, silver iodide is quite soluble, and it is 
possible to fix out emulsions in solutions of potassium iodide. But if the silver 
image is imbedded so that it is completely surrounded by gelatin that has 
been hardened, the formation of the iodide will take place with the accom- 
paniment of adsorbed potassium iodide. The new complex has a great molecu- 
lar weight, and will move very sluggishly through the gelatin, so that in the 
short time which is required for the conversion of the image into the complex, 
it will have moved but an infinitesimal amount. That is the reason for the 
relatively fair definition of a silver-iodide, potassium-iodide, complex mordant. 
The same is true for the formation of the other complex mordant images of 
the Christensen type, which use thiourea and thiourea derivatives for the 
formation of the complex. But we can leave these questions for a later dis- 
cussion. They are being mentioned here only to illustrate the influence that 
high molecular weight has upon conferring substantive properties to any 
molecule. The I.G. research workers have sought to "dye" the gelatin with 
the coupling reagent. This is a transparent and colorless dye, which will not 
have a tendency to wander about when the gelatin layer is made wet during 
the processing stages. 

But this is not the only problem that is involved. The presence of very 
heavy groups in the coupler molecule has a tendency to yield color shades that 
are distinctly duller than those obtained if the substitutions are left out. It 
becomes extremely important, therefore, not to lose any brilliance in the 



ANSCO COLOR AND AGFACOLOR 131 

remaining stages of the processing technique. One trouble spot in that con- 
nection may be overlapping color sensitivities in the three emulsion layers. 
This danger was recognized early, and considerable efforts were made to over- 
come it. It became necessary to develop new sensitizing agents whose spectral 
cuts would be sharp, and mutually exclusive, at least when used in connection 
with the proper niters. The red sensitizer must have a pronounced gap in the 
blue-green region of the spectrum. Pinacyanol fails completely in this respect, 
since it has a considerable activity in the green. Sensitizers fulfilling these 
conditions are described in English patents 383486, 400951, 402521, 410481, 
427887, 432969, 442160, 480778, 483548, and 486005. The exact requirements 
for the sensitivity of the various zones are described in greater detail in English 
patent 501 190. The blue filter layer should have a sensitivity ranging from 
400 m/x to 480 m/x. This indicates a considerably sharper cut than is generally 
used in three-color work, where the blue cut usually is allowed to extend to 
510 or even to 520 m/x. Since there is a yellow filter between the blue and the 
other two layers, the sensitivity of the green below 500 is of no account. Light 
whose wavelength is less than that, will not be able to reach the green-sensitive 
zone. Its sensitivity should not extend beyond 610 m/i, and the maximum 
should be at 540. The red-sensitive layer should have its maximum lying be- 
tween 580 and 630 m/x. This would indicate that Agfacolor has a range ex- 
tending from 400 mfj, to approximately 650, and the sensitivities should sepa- 
rate from each other at 480 m/x and at 570 mju. This again is quite different 
from standard procedure, which separates the three zones at 500 and 600 m/z. 

The complete process is described in English patent 481501, issued on 
March 7, 1938. The monopack layers are formed as follows. Upon the base 
is coated a red-sensitive emulsion layer to which has been added i-N-stearoyl- 
4-N-(i , -hydroxy-2 / -naphthoyl)-phenylenediamine-sulphonicacid,in the form of 
the sodium salt. It is interesting to note that this compound contains at least 
three conditions that would make it substantive to gelatin. One amino group 
of the paraphenylenediamine has been converted into a stearoyl-amide which, 
as has been pointed out, gives the molecule fastness to diffusion by the inclusion 
of an 18-carbon chain, hydrocarbon residue. The entire molecule is given an 
acid character by the inclusion of a sulphonic acid group. This makes it 
substantive to gelatin by virtue of its ability to form gelatin sulphonates. The 
third condition is the substitution of the phenylenediamine residue into the 
alpha-hydroxy-naphthoic-acid molecule, for it is the naphthol which combines 
with the oxidized developer to form the cyan dye. Any chance that the 
phenylenediamine molecule may have had to unite with the developer was 
destroyed when each amino group was converted into an amide, one by con- 
version into stearoyl amide, the other by conversion into naphthoyl amide. 
This explains the reason for the inclusion of an acid group into the naphthol 
molecule., 

The second, or green-sensitive emulsion layer contains the sodium salt of 
i-(5 / -sulpho-3 , -stearoyl-aminophenyl)-pyrazolone. Here again the substantive 
character of the pyrazolone molecule is taken care of, at least three times. 



132 HISTORY OF COLOR PHOTOGRAPHY 

First, by the inclusion of the 18-carbon chain stearoyl residue; second, by the 
inclusion of a sulphonic acid radical; and third, by the substitution within 
the pyrazolone molecule of an amino-phenyl group. The coupling properties 
of this last group are destroyed by masking the amine group with the stearoyl 
residue. The blue-sensitive layer contains the sodium salt of meta-stearoyl- 
amino-benzoyl-acetanilide-para'-carboxylic acid. The molecule is made sub- 
stantive first by the inclusion of the very heavy stearoyl residue, second by 
the inclusion of the acid carboxyl group which would allow the formation of a 
gelatin-carboxylate salt, and third by the presence of an amino-benzoyl group 
whose coupler properties have been masked when the amine group has been 
converted into a stearoyl-amide. It is to be noted that besides containing 
specific groups which yield substantive character, the molecules are all very 
heavy ones so that even if conditions are formed which would prevent their 
union with the gelatin, they would have a very limited range about which they 
could roam in the time during which they are sufficiently wet to be subject to 
diffusion. The cyan coupler present in the red-sensitive layer has a molecular 
weight of 613. The magenta coupler in the green-sensitive layer has a molecu- 
lar weight of 543, while the yellow coupler present in the third layer has a 
molecular weight of 582. These are really giant molecules, which could hardly 
exist in solution except in colloidal state. 

Despite all the efforts which have been made to find new agents which would 
yield color-coupled dyes, it is interesting to note that all the cyan dyes are 
formed either from the hydroxy benzenes, or from the hydroxy naphthalines. 
This is true not only of all the coupling agents disclosed by Ansco and the I.G., 
but also of those disclosed by the members of the Eastman Kodak staff. And 
in the very early descriptions of Fischer and Siegrist, these substances were used 
for the formation of the cyan image. All that has been done in the generation 
that has passed since that time, is to tack on certain side-chain substitutions 
into the molecule. And Fischer and Siegrist outlined in some detail what the 
effect of substitutions .would be. The magenta couplers fall into two classes. 
First there is pyrazolone and the pyrazolone derivatives. These couplers 
appear to be preferred by Ansco and the I.G. Then there is benzyl cyanide 
and its derivatives. These appear to be favored by Eastman. The yellows, 
without exception, are all compounds which contain the central grouping 
— CO — CH2 — CO — . The only difference between the very many compounds 
that have been proposed, is in the nature of the groups which round out the 
molecule. Two such groups are needed to form the compound, and these can 
be anything from a simple hydrocarbon to a complicated substitution. 

After exposure, the pack is developed for fifteen minutes in an amidol depth- 
developer. 

Amidol 5 parts 

Sodium sulphite 50 parts 

Potassium bromide 1 part 

Water to 1000 parts 



ANSCO COLOR AND AGFACOLOR 133 

A ten-minute wash removes excess developer. The negative image is next re- 
moved by treatment with a solution containing half of one per cent of chromic 
acid, or with 

Potassium dichromate 4 parts 

Sulphuric acid 2\ parts 

Water to 1000 parts 

The pack is washed again for a period of ten minutes. The treatment with acid 
dichromate or chromic acid desensitizes the silver halide salts remaining in the 
emulsion, so that it becomes necessary to restore the sensitivity by treatment 
for two minutes with a three per cent solution of sodium sulphite. After a 
further five-minute wash, the film is exposed to light until all the remaining 
grains become developable. The second development is carried out in the 
amidol depth-developer noted above. Instead of giving the pack an exposure 
to light to make the remaining grains developable, it is possible to treat it 
chemically with a fogging agent such as thiourea and thiourea derivatives. 
This is suggested in English patent 489845, but of course similar ideas have 
been suggested ever since reversal development became an established practice. 
The Eastman Kodak Company, in its procedure, mentions the use of these 
agents, as well as the use of dyes like methylene blue, etc. Thiourea was 
added to developers as far back as 1910 or thereabouts, when it was suggested 
to develop a positive image directly after the first exposure. Even at that 
time its action as a fogging agent of great power, was known. There is, there- 
fore, but little novelty in this specific patent. 

It may be wondered just why the second development was not carried out 
with phenylenediamine, so that conversion to color would take place immedi- 
ately. That would have made unnecessary the patent which is being described 
now. But a more serious reason is that it may be desirable to see the positive 
in the form of a black-and-white silver image, and to note any corrections that 
may be necessary. Reduction, or intensification by means of silver salts, would 
then be possible. Since for every four atoms of silver there is formed one 
molecule of dye, the change in silver would change the dye content. But the 
only changes that would be possible would be overall changes in contrast, for 
it would be rather a difficult matter to treat the silver image, say in the bottom 
layer, without at the same time treating the silver images in the other two 
layers. After all manipulations have been carried out, the silver is converted 
into silver chloride by treatment with a bleach of the wash-off relief type. 



Chromic acid 


5 parts 


Sodium chloride 


50 parts 


Water to 


1000 parts 


or 
Potassium dichromate 


4 parts 


Hydrochloric acid 


2 \ parts 


* Sodium chloride 


20 parts 


Water to 


1000 parts 



134 HISTORY OF COLOR PHOTOGRAPHY 

The silver chloride is now ready for color-development. This is accomplished 
by treatment of the reformed salt with a paraphenylenediamine developer 
made as follows: 

Sodium sulphite J part 

Di-ethylparaphenylenediamine 2§ parts 

Potassium bromide 2§ parts 

Potassium carbonate 30 parts 

Water to 1000 parts 

The time of development is twenty-five minutes. Since the coupling agents 
are already present in the film, it is not necessary to include them in the de- 
veloper formula. Therefore the same solution is used to develop the three 
layers, and there is formed a magenta image in the green-sensitive layer (by 
condensation of the oxidized developer with the agent present in that layer), 
a cyan image in the red-sensitive layer (by condensation with the agent present 
there), and a yellow image in the blue-sensitive layer (by reaction with the 
substance in that zone). It is very important that the wash after this step be 
exceptionally good, so that no more developer is present in the film when it is 
subjected to the action of a 5 per cent ferricyanide solution to which has been 
added 5 per cent sodium chloride. Otherwise a general color stain will result. 
Hypo removes the silver chloride formed in the last step, and after a final wash 
the positive transparency is complete. 

As outlined above, this process differs materially from the early Koda- 
chrome. There a complicated system of processing was required before the 
three latent images were converted into their proper colors. Here all the 
complications have been removed, and only four simple steps are required to 
obtain the final result. The first is an ordinary development in a depth de- 
veloper that is well known to photographic technicians. Second there is the 
removal of the silver image by treatment with a silver solvent. This step also 
has been well exploited by the average color technician who had previous ex- 
perience with the older screen plate processes (cf. chapter on Screen Plates). 
The third step is redevelopment with a special developer, whose composition 
is well known. The last step consists in the removal of the silver image, and 
this can be done either in a single bath (Farmer's reducer) or in two stages, as 
described above. 



CHAPTER 12 
SCREEN PLATES 



Th, 



lHE integral methods for the preparation of color photographs are not 
confined to the monopacks. Ducos du Hauron suggested a much simpler 
method almost fifty years before the first monopack was even conceived. This 
he disclosed in his French patent 83061 issued in 1868, and developed further in 
a series of papers which were later collected and published in book form, "Les 
Couleurs en Photographie," in 1869. If a base be completely covered with 
alternating lines of red, blue, and yellow, so that no empty spaces lie between 
them, they will not be seen at normal viewing distance as individual lines, if 
they are sufficiently fine. They will be seen as a gray, the hue that would 
result from the fusion of the three secondary colors. This effect can best be 
visualized by an examination of a halftone reproduction of a picture in news- 
papers and in magazines. A newspaper picture is usually reproduced by 
means of 40 to 60 lines to the inch. At normal viewing distance the individual 
units could just be seen, giving rise to a picture with coarse detail. A maga- 
zine illustration utilizes a minimum of 120 lines to the inch, and these cannot 
be seen individually except under a small-power glass. The picture under 
these' conditions appears smooth and continuous. To make a motion picture 
equally smooth, it would require a minimum of 500 lines to the inch. 

If the transmission through one of the lines were blocked out, the light that 
would result would be clue to the fusion of but two primaries, and it would no 
longer be white or gray. Thus, in order to obtain a red color, it is sufficient to 
block out the blue and yellow lines. To obtain blue, the other two are blocked. 
It is possible to obtain any combination of colors by merely partially blocking 
out each of the lines, to a different extent. Black would be obtained by com- 
pletely blocking out all three of the lines. 

In making such a pattern, du Hauron pointed out that it is necessary that 
the fusion of the three lines, when none of them are blocked, give rise to a pure 
white. This was later described as the "first black condition" by Mees and 
Pledge. Then the uniform blocking of the three elements will give rise to 
neutral grays. If the elements of the screen are very coarse, it is possible to 
manually block out each primary color to the desired degree. But in the case 
where complete fusion is desired, the elementary cells would be too fine to 
enable an operator to deposit light-retarders behind each unit. Du Hauron 
realized, however, that the rapidly growing art of photography enabled a 
person to make blocking-out a routine and automatic procedure. All that was 

13s 



136 HISTORY OF COLOR PHOTOGRAPHY 

necessary was to put a light-sensitive emulsion over the screen and allow the 
light to fall on the emulsion only after passage through the screen elements. 
The emulsion must be such that it can be processed directly to a positive, so 
that the different light intensities become immediately transformed into equiva- 
lent transparencies. 

To make the procedure possible, the emulsion had to be sensitive to all the 
rays of the spectrum. If it was more sensitive to one group of rays than to 
another, it was possible to place in front of the areas occupied by the more 
sensitive colors, light retarders to such a degree that equivalent speeds resulted. 
This was later known as the "second black condition." It is equivalent to 
placing neutral filters in front of the color elements. A more useful method for 
the balancing of the emulsion speeds to the three primary colors, was to vary 
the relative areas covered by each of the color elements. 

Up to this pomt du Hauron merely outlined, in general terms, a new method 
that could be used to make color photographs. He did not disclose how such a 
screen could be made until a considerable time had elapsed. It was not until 
1897 that Alcide Ducos du Hauron, who was a brother of L. Ducos du Hauron, 
published a book, "La Triplice photographique et lTmprimerie," in which he 
outlined two general procedures for the preparation of screens. These dis- 
closures have great historical value, since they predate all other patents that 
disclose the formation of screens. The first method stated that it would be 
possible to rule the lines mechanically, using an obvious procedure. The second 
method is one which has been patented and repatented a countless number of 
times, not only in the exact form outlined by du Hauron, but with all possible 
variations. And yet no successful screen plate utilizing this procedure has 
been marketed. 

The scheme utilized the action of light upon dichromated gelatin which had 
been previously dyed a primary red, and which was coated upon a transparent 
base. It was exposed under a negative which consisted of a series of opaque 
and transparent lines, the width of the opaque ones being twice those of the 
others. Upon exposure, the gelatin receiving the light becomes converted into 
insoluble form. Hot water removes the gelatin protected from the light by 
the opaque lines. This leaves a series of fine red lines upon the base, which 
cover one third of the total area. The coloring matter is firmly fixed within 
the gelatin by a further treatment with mordants and tanning agents. After 
being thoroughly set and insulated, the base is recoated with sensitized gelatin, 
and this time dyed green. It is again exposed through a negative, but now the 
lines run at right angles to those previously formed, and the opaque regions 
are equal in width to the transparent ones. The sensitized gelatin is exposed 
through the red lines, so that these act also as opaque lines. By this means, 
one-half of the areas between the red lines become converted into green areas. 
The remaining gelatin is removed by hot water, the green areas treated to fix 
the coloring matter in a permanent form, and the plate recoated again, but 
with sensitized gelatin dyed blue. The last exposure is made only through 



SCREEN PLATES 137 

the base of the plate, the red lines and green squares acting as a negative. 
Since dichromated gelatin is insensitive to red and green light, only the regions 
free of them will be affected. In this manner the entire surface of the plate 
becomes covered with the color elements, no clear spaces existing between 
them, thus satisfying the fundamental conditions. 

From the above it is seen that du Hauron had a very good conception of the 
procedure involved in screen plates. One curious fact stands out. In his very 
first disclosures, du Hauron speaks of the line elements being colored red, 
yellow, and blue. Thirty years later he makes the colors red,, green, and blue. 
The effect of loose terminology is quite apparent here. The blue and red in the 
first set cannot be the same as the blue and red of the last. Yellow is a 
minusing primary, and in the set where it forms one of the colors, a magenta 
rather than a red, and a cyan rather than a blue, are required. Color pho- 
tographers have fallen into the very bad practice of calling the "magenta" a 
red, and the "cyan" a blue. These misnomers arose from the fact that the 
printing inks available in the early days were poor. Even today, the actual 
color of the minus-green printing ink is much more nearly a red than a magenta, 
and the minus-red printing ink is almost a blue. We must assume, from the 
neutrality conditions that he himself set up, that in the first set du Hauron 
actually meant to use magenta, yellow, and cyan colored elements. When 
such dots lie in juxtaposition it is possible to obtain a neutral, but if the 
magenta be replaced by a red, and the cyan by a blue, it will be impossible 
to obtain a neutral gray, but rather a highly diluted and unsaturated red. 

The effect of photographing with magenta, cyan, and yellow elements, in- 
stead of the normal set, would be to dilute the colors with white. The image 
behind a yellow line, for instance, would contain a silver deposit that is due to 
the red and the green primaries. The red densities would be deposited also 
behind the magenta element, this time together with the blue. Thus there is 
no clear-cut separation of the primary colors. A bright red color would yield 
deposits behind the yellow and the magenta lines, and no deposit behind the 
cyan. Upon conversion into a positive, the transmission through the cyan 
element would be zero, and through the yellow and magenta elements 100 per 
cent. But yellow contains equivalent portions of red and green, while magenta 
contains equivalent portions of red and blue. The fused beam would contain 
equivalent amounts of the blue and green primaries, and double that quantity 
of red. One unit of red will combine with the blue and green to yield a white. 
The final color will be red diluted with an equal quantity of white. It is sig- 
nificant that in his latest reference to the subject du Hauron no longer made 
that error, but utilized red, green, and blue. 

In the preparation of a screened surface, several problems confront the 
technician. First of all, the spectral characteristics of the individual line ele- 
ments must be considered. The fact that the process is an additive one, should 
fix these rigidly, but the early workers in color photography were still under 
the influence of the psychologists, and considerable debate was occasioned by 



138 HISTORY OF COLOR PHOTOGRAPHY 

the lack of precise knowledge in this respect. The physicists were strangely 
silent. Perhaps the apparent success of the Young-Helmholtz-Maxwell ideas 
concerning color mixtures throttled all other approaches in the field of color 
reproduction, just as it did in the field of color vision. It was only in the 
middle third of the twentieth century that the physical and biochemists began 
to challenge the other approaches to the study of color and vision. Unfortu- 
nately, color photography could not wait until a satisfactory solution to this 
fundamental problem had been found. 

The first experimenters utilized the data on color mixing that had been 
gathered by Maxwell, Koenig, and others. Prof. J. Joly {Brit. J. Phot., Vol. 42 
(1895), p. 774) rejected the Maxwell curves in favor of those by Koenig. His 
blue element approximated in color monochromatic light of wavelength 455 m/x. 
The red color was approximately that corresponding to a wavelength of 600 m/x, 
and the green corresponded to a wavelength of 550 m/x. Actual measurements 
made on a Joly plate by Kaiserling {Phot. Mitt., Vol. 35 (1898), p. 273; Vol. 36 
(1899) p. 8, 14, 35, 46, 65) indicated that the red transmission began at 550 and 
extended to 670 with a maximum at between 580 and 600. The green began 
at 470 and extended to 570 with a maximum between 515 and 530. The blue 
ranged from 430 to 520 and had a maximum at 460 to 480. Thus, while precise 
specifications were laid down and theories accepted or rejected if they did not 
correspond to these specifications, the practice was really very much the op- 
posite. The red overlap into the green appears to have been almost to the 
point of the maximum green transmission, and a very small gap existed between 
the limits of the blue and the red primaries, a gap sufficiently small to make 
these colors better suited for a two-color separation than for a three. The green 
transmission overlapped the blue up to the point of maximum blue trans- 
mission. The blue affected the green in a like manner. Such color separations 
may give plates with high speed ratings, and perhaps yield original color 
photographs that would appear quite pleasant, but the separations made from 
them would yield prints of very poor quality, unless color corrections were 
applied. Abney ("The Scientific Requirements of Colour Photography," 
1897), also insisted that the elements correspond to the three fundamental color 
sensations, but because this would entail the use of a green highly diluted with 
white, he made a compromise upon this score. His green corresponded to a 
color with a wavelength of 527 m/x, which was but slightly diluted with white. 
The red, which according to his theory should correspond to a wavelength of 
671, was also a compromise and actually corresponded to light with a wave- 
length of 656. The blue corresponded to a wavelength of 460 m/x. 

Mees {Brit. J. Phot., Vol. 55 (1908), p. 41), and Mees and Pledge {Phot. /., 
Vol. 50 (1910), p. 197), made an exhaustive study of the entire subject. These 
men concluded that the practical requirements demanded that the red have a 
transmission which begins at 590; the green a transmission ranging from 490 
to 590; and the blue, a transmission up to 500 m/x. These are practically 
mutually exclusive filters that correspond very closely to the physical require- 



SCREEN PLATES 139 

ments for color analysis (cf. chapter on Objective Color). The F, N, and C4 
niters, which are the Wratten filters Nos. 29, 61, and 49, approximate these 
transmissions. As was indicated previously, their use would entail a tremen- 
dous loss of light efficiency, due to the inherent inefficacies of the dyes, so a 
compromise has been made in the use of the somewhat overlapping set, A, 
B and C5. -The condition that the fused elements should yield a pure white, 
the first black condition previously pointed out by du Hauron, must be met 
by a variation in the ratios of the areas covered by the respective elements, 
rather than by change in the densities of the colors. These must give complete 
cutouts. 

Also very important in the manufacture of a screen, are the questions relating 
to screen pattern, and the size of the elements. W. Scheffer (Brit. J. Phot., 
Vol. 56 (1909), Col. Supp., p. 62, 70) tackled this last by a study of the relation- 
ship between object size and the resolving power of the eye. He found that 
black-and-white objects will be seen as distinct items if their period is 1/1000 
of their distance from the eye. The period is the distance between two lines of 
like character in a screen containing opaque and transparent lines of equal 
width. E. J. Wall ("History of Three-Color Photography" (1925), p. 462) 
pointed out that color does not follow this rule. Color elements fuse much 
sooner. The size under these conditions must be 1/600 of the distance of the 
object from the eye. With the irregular screen, the possibility of clumping 
further modifies these values. Mees and Pledge finally conclude that in geo- 
metrical screens the units should have a size varying between 1/300 to 1/600 
of an inch. Irregular screens should have elements whose sizes fall between 
1/900 to 1/1200 of an inch. This, of course, relates to screens to be used for 
normal viewing, and not for projection. The Autochrome plate contained 
colored starch grains whose average size was 0.015 mm. This corresponded to 
0.0006 inch, or 1666 to the inch. 

The first black condition demands that the entire surface of the plate be 
covered by the screen elements, and that no part of the area be free from color. 
If the screen is formed by mechanical dyeing or inking, there is danger that the 
color in one element will run over into an adjacent one. At any rate, the line 
formed by the juncture of two elements will contain both dyes, and to all in- 
tents and purposes, will be black. As these lines will have some dimension, 
in order not to lose too much light efficiency their length must be kept to an 
absolute minimum. C.W. Viper (Brit. J.Phot.,Vo\. 50(1909), Col. Supp., p. 84), 
made a study of this phase of the subject. His results indicated that the length 
of the lines formed by the junction of two screen elements will be a minimum 
where the screen is composed of parallel lines. A close second to this is the 
case, used by Dufaycolor, where the elements consist of parallel lines of one 
color interposed with squares of the other two. 

Finally, there is the problem of putting a panchromatic emulsion on top of 
the screen. To make the emulsion stick, a substratum must first be coated 
directly over the screen. This serves two purposes. First it insulates the 



i 4 o HISTORY OF COLOR PHOTOGRAPHY 

screen from the photographic emulsion and prevents the action of water or 
photographic processing solutions upon the ingredients in the screen elements. 
Then it serves to bind the emulsion layer to the plate or film. The thickness 
of this substratum must be as low as possible so that too great a separation 
will not exist between the screen and the photographic emulsion. The emulsion 
must be as rapid as possible, for there already is lost a considerable amount of 
light by absorption by the screen elements, etc. Therefore there is a limit to 
the fineness of the grains. On the other hand, they cannot be too coarse, for then 
too few of the emulsion grains would be located behind any one screen element 
to make it an average sample of the entire emulsion. Weird and uncertain 
results would be obtained. This problem will be discussed in detail a little later. 

The emulsion must be panchromatized in such a manner that when the ex- 
posure is made through the screen elements, equal times of exposure and de- 
velopment will yield equal densities behind each element in those regions which 
correspond to whites and grays. This is known as the second black condition. 
It is not a very easy matter to accomplish such a result, because it is difficult 
to make two batches of emulsion having identical characteristics. One needs 
only to record the variations in the values of the filter factors of a given type 
of emulsion from coatings that are six months or a year apart, to determine 
this fact. A variation of the relative areas of the screen elements would prob- 
ably be sufficient to maintain this condition, but it is a question whether a shift 
in these ratios made to fulfill the second black condition would not, at the same 
time, disturb the first black condition. 

It has been suggested that the emulsion grains behind each screen element be 
sensitized only for the color which that element transmits. This can be done 
by adsorbing the sensitizer upon the colored screen element, and coating a blind 
emulsion directly upon this. Several problems are at once introduced. First 
there is no guarantee that a sufficient sensitization would be had by this 
technique, for only a few layers of grains would be affected. Secondly there is 
the problem of coating an emulsion directly on top of the screen elements with 
no substratum to bind the two together. Frilling and separation of the two are 
very likely to occur under such conditions. Finally there would be the problem 
of choosing the colors for the screen elements, so that they would not only ad- 
sorb sufficient sensitizing dyes to act upon the emulsion, but would also be 
completely impervious to the action of the processing solutions, of which one is 
acidified dichromate or permanganate, powerful oxidizing agents, especially 
toward the organic dyes. 

A much better method has been disclosed by the I.G. (Eng. P. 4591 19). A 
basic emulsion is divided into three portions, and each portion is sensitized 
to a different primary. Portions of each emulsion are then tested behind 
filters that are identical with the ones in the screen plate itself. By this means 
the relative sensitivities of the emulsion to the three primaries are determined. 
The grains are then mixed thoroughly in the proportions required to yield 
exact balance. ' 



SCREEN PLATES 141. 

The emulsion should be as fine-grained as it is possible to use with a given 
size screen element. The greater the area covered by each unit, the coarser 
can be the emulsion. It should be recalled that an emulsion is not a dispersion 
of homogeneous light-sensitive particles. The sensitivities of the individual 
grains vary tremendously. It is only the fine-grained, very slow, and extremely 
contrasty emulsions that approach any degree of uniformity. An emulsion 
that has considerable latitude and wide covering range is possible only because 
of the existence of a large variety of sensitivities among the individual grains. 
It has been pointed out that in extremely fast emulsions, only five per cent of 
the grains are sufficiently sensitive to become developable by the absorption of 
but a single quantum of light. But it requires 200 quanta to make the average 
grain in the same emulsion developable. Therefore, there must be at least 
several hundred types of grains present, and the emulsion has properties which 
are a statistical average of the properties of each of these varieties. It is 
essential, therefore, in order to achieve uniformity, that behind each unit of 
screen there be present a sufficient number of grains to be representative of 
the emulsion as a whole. This number must be counted not in hundreds, but 
in thousands and tens of thousands. And since it is desirable to keep the image 
as close as possible to the screen itself, each grain layer of emulsion must have 
this number behind each color unit. 

Most of our information concerning emulsions comes from that brilliant 
team, Carrol and Hubbard, who did their work at the Bureau of Standards, in 
Washington, D.C. They used an experimental emulsion that represented a 
normal speed at the time they made their disclosures (Bu. Stand. J. of Res., 
Vol. 7 (193 1) p. 229). The projective areas of the grains in this emulsion ranged 
from 0.36 to 1. 00 square fJL, so that each grain had a linear dimension ranging 
from 0.6 to 1.0 /x. These values have been corroborated by others, as repre- 
senting the grain size in the average negative emulsion. The grains are dis- 
persed in gelatin, and between any two grains there lies a wall of gelatin. The 
two substances, gelatin and silver halide, are present in equal amounts by 
weight, but gelatin is much less dense than silver, so that the volume of gelatin 
is much greater than that of halide. But we can assume exact equivalence. 
Hence the average distance between grain centers will vary from 1 to 2 fx, so 
that there will be from 500 to 1000 grains to the linear millimeter of emulsion 
surface, or from one-half to one grain per /x. 

The Autochrome plate contained color dements that were about 1666 to the 
inch. It has sometimes been claimed to have 2000. Taking the last value, 
this means that the cell has a linear dimension of 12^ /x, and an area of 156 
square ju. In order for this space to contain 1000 grains, there would have to 
be 30 grains in 12-3- /x, or each grain would have to be at most 0.4 ft in size. 
This explains to a large extent why the Autochrome plate was sixty times as 
slow as the normal negative emulsion of that period. The Agfa plate was 
exactly twice as fast as the Autochrome. 

The Dufaycolor film contains color elements with a cross section of 0.002 



142 HISTORY OF COLOR PHOTOGRAPHY 

inch or 50 ju. Therefore between 625 and 2500 grains can lie beneath each 
screen element, per layer of emulsion. There are approximately thirty layers 
so that a total ranging from 20,000 to 75,000 grains will lie in each unit. This 
represents a good average for the emulsion as a whole, and the individual layers 
will also be fairly representative. Under these conditions, there is hope for 
latitude and range. The popularity achieved by this material can no doubt be 
ascribed in large measure to the fact that fast emulsions can be used in the 
preparation of Duf ay color film, whereas screen plates with finer screen ele- 
ments must use finer-grained, slower, and more contrasty emulsions. Unless 
our emulsion makers can make superfine emulsions, with high speeds, it is 
doubtful if screens with more than 500 elements to the linear inch can be made. 
The speed of an emulsion is to a large extent determined by the intensification 
factor of the developer. This term is defined as the ratio of the number of 
atoms of silver formed to the number of quanta absorbed by the grain to make 
it developable. Under present conditions (1943), the fast emulsions have 
values of 10,000,000. If the grain is made materially smaller, the number of 
silver atoms in the grain becomes smaller. The problem is therefore to make 
the grains faster by an amount greater than the size decrease. Considerable 
leeway is had in this respect, for in the fast emulsions, 200 quanta of light must 
be absorbed before the average grain becomes developable. Theoretically, 
only one quantum should be required. 

Many ingenious schemes have been formulated for the preparation of the 
screen. These can be classified into a number of groups, the members of each 
group being related by the technique used. The groups are as follows: 

1. Dusting-On Methods 

2. Woven Fabrics 

3. Blocks 

4. Photographic Printing 

5. Photo-Mechanical Printing 

6. Ruled Lines 

7. Resists 

8. Miscellaneous 

A critical discussion of these procedures will be given in the following chapter. 
Here will be offered only a generalized description. 

In the dusting-on schemes, finely powdered colors are dusted over a tacky 
surface. A gentle blast of air removes the particles that are not directly adher- 
ing to the surface, and thus only one layer of the grains becomes attached. If 
the powders are properly mixed, each red grain will be surrounded by blue and 
green grains, and vice versa. In some schemes the dry powders are first dis- 
tributed in a colloidal medium like gelatin, nitro cellulose, gums, etc., and this 
dispersion is coated upon the base in a thickness just sufficient to yield a single 
layer of colored particles. Obviously this is but an improved form of the dust- 
ing-on methods. All kinds of substances have been suggested for a source of 



SCREEN PLATES 143 

the powders, including glass, nitro-cellulose, gums, varnishes, gelatin, pollen, 
microbes, bacteria, starch, oil drops, etc. In general, the substances are dyed, 
then ground or atomized to powder form, mixed to fulfill the first black con- 
dition, and coated upon the base. 

The rapid advances made by the textile industry in the making of extremely 
fine fibers soon brought the suggestion that such fibers, colored properly, could 
be woven into fabrics. The colored fibers would all be woven in one direction, 
while the fibers that would weave in the other direction to hold the fabric to- 
gether, would be colorless and transparent. The artificial silk fibers, rayons, 
etc., were especially suitable for this purpose. The fibers could be drawn very 
fine, to the order of a few thousandths of an inch, hence screens suitable for 
normal viewing could be made. 

One method for the preparation of the screens suggests itself naturally. Thus 
sections of nitrocellulose, gelatin, etc., are dyed properly, then placed one on 
top of the other until a solid block is built. A cross section of this block will 
show a series of lines alternating in the three colors. The sections can be cast as 
thin as desired, even less than 0.0001 inch in thickness. Any degree of fineness 
could therefore be obtained. After formation, the block is cut across the 
layers by means of a microtome, giving sheets of screens suitable for immediate 
use. It is also possible to cross section a block of this type, reform the block, 
and cut it again from a different direction. Instead of lines, there will be 
obtained a checkerboard mosaic. Another possibility is to convert the block 
into a cylinder, and section this spirally, to yield a long continuous film of 
screen. It is not a difficult matter to cut sections 36 by 72 inches as thin as 
0.005 mcn - Smaller sheets can be cut still thinner. Such film sections can be 
used not only as a screen, but also as the base upon which the emulsion is cast. 

The most obvious method for the formation of the screen is to rule the 
colored lines upon a base. This presents some technological problems, since it 
is essential that one element just touch its neighbor with a minimum of over- 
lapping. Only slightly less difficult are the methods by which the screen is 
printed by photo-mechanical methods, such as letter press, lithography, or 
offset. Here, again, registry problems arise that are considerably more diffi- 
cult than in ordinary printing. In the present instance it is required to print 
screens with a minimum of several hundred elements to the inch, while in 
normal printing the screens have approximately 120 lines to the inch. 

These difficulties are not met with when photographic printing is attempted. 
The more successful methods utilized dichromated gelatin. This was coated 
upon a base, and exposed under a black-and-white screen. The exposed gelatin 
became tanned and would absorb only certain types of colors. Or a pinatype 
dye could be put into the gelatin before the exposure, in which case it would 
wash out only from the non-exposed portions. Two procedures are open. The 
soft gelatin could be removed by hot water, and the plate recoated, this time 
with dichromated gelatin dyed another color. The new exposure is made at 
right angles to the first. This fills up half the remaining spaces with a dyed 



144 HISTORY OF COLOR PHOTOGRAPHY 

gelatin. The procedure is finally repeated to put on the third color. The other 
possibility is to take advantage of the fact that certain dyes will stain only 
tanned gelatin, while other dyes will sjain only soft gelatin. The plate, after 
the first exposure, contains a series of lines of tanned gelatin separated from 
each other by areas of soft gelatin. These could be dyed different colors, and 
the untanned gelatin again sensitized and exposed at right angles to, the first. 
The dye is washed out of the soft gelatin, and this is dyed with the third color. 

Anything that can be accomplished by the action of light on dichromate, 
could also be accomplished by means of a silver image. It is possible therefore, 
to coat a plate with a silver emulsion, expose it to a line screen, and treat the 
silver to give lines of tanned gelatin interspersed with areas of soft gelatin. 
The further procedure could be as above. It is also possible to convert the 
silver image into a primary color, by toning methods or by coupler develop- 
ment. This would leave silver bromide in the areas between the toned lines, 
which could be treated in any desired fashion to yield the other two colors. 

The most successful of all screen processes, Dufaycolor, utilizes the resist 
methods for the formation of the screen. Here a fatty ink is printed by mechan- 
ical means upon the surface of the dyed film. The areas between the ink lines 
are subjected to a solution which dissolves out, or which destroys the coloring 
matter. These areas are next dyed a different color, the resist lines removed, 
and a new set printed at right angles to the first. Again the coloring matter is 
removed, and the third dye placed in the now colorless spaces. A variation of 
this would be to print, mechanically, a waterproof substance upon a photo- 
graphic emulsion. The areas between the resists could now be treated with 
coupler developers. The resist is then removed and a new one put on, covering 
half of the remaining unreacted space; or, only half of the resist could be re- 
moved. The newly exposed emulsion areas can be again developed to a differ- 
ent color. The removal of the resist now exposes the last third which can be 
converted to the third primary. Instead of a series of color development, it 
is possible to convert the silver halide into silver iodide or silver iodide-thiourea 
complexes which are completely transparent and highly absorbent of basic 
dyes. Or, the photographic emulsion, before the resist is put on, can be con- 
verted into a silver iodide complex and dyed uniformly one color. A resist 
could be printed upon it, and the dye halide complex destroyed by the action 
of acid permanganate. Treatment with sulphite removes the reduced manga- 
nous salts, and treatment with iodide reconstructs the silver-iodide complex. 
This is now dyed with another color, and the entire procedure repeated. These 
last few ideas are discussed at this point rather than in the next chapter be- 
cause they are not disclosed in the patent literature, and are only hypothetical 
processes. 

The efficiencies of screen plates are not very high. This is due mainly to the 
fact that the individual screen elements transmit but a fraction of all the light 
falling upon them. Measurements by Mees and Pledge made upon the plates 
commercially available in 1910, show that only 20 per cent of the light was 



SCREEN PLATES 145 

transmitted by the most efficient of them all, the Dufay plate. This was not 
the final Dufaycolor film, but an early edition of the later product. Other color 
plates of the time transmitted as follows: Autochrome 7f per cent; Thames 
(Finlay) 12 per cent; and Omnicolore 15 per cent. It is seen, therefore, that 
the Autochrome plate would have a factor of 14, were it able to use the normal 
emulsion. But its factor was 60. 

The introduction of Kodachrome, Ansco Color and Agfacolor dealt a severe 
blow to the popularity of the screen plates. There are several basic advantages 
to the monopack schemes which would explain their immediate acceptance. 
First of all, the lens speeds of the newer materials were at least equal to, 
and in all probability greater than, those obtained with screen plates. The 
popular demand has ever been for speed and more speed. Secondly, the 
transparencies obtained by their use are continuous in tone, so that enlarge- 
ment of four diameters or more can be made, whereas with the use of a screen 
plate, the enlargement is limited to not more than two diameters. In the 
matter of making separations from the transparencies, the use of monopacks 
is to be preferred, since the overlaps between the "primary colors are much less. 
There is one field, however, that appears to be ideally suited for the screen 
plate. This is the use of a screened bipack in which the front element is a 
screen plate which registers two of the primaries, and the rear element registers 
the third. This procedure was first disclosed by Ives (Eng. P. 112769; Fr. P. 
487529; U.S.P. 1268847). The screen units in the front element were colored 
yellow and magenta, and these were coated with an emulsion sensitive only to 
the blue and green primaries. The blue densities would be registered only 
behind the magenta elements, while the green densities would register only be- 
hind the yellow lines. By the use of filters, it therefore becomes possible to 
copy the images behind these lines upon separate plates. Both the yellow and 
the magenta lines transmit the red rays with great efficiency. Since the front 
element is completely insensitive to this section of the spectrum, it will register 
only upon the rear element, and in the form of a continuous tone negative. 

The making of a two-color screen is very much less complicated than the 
making of a three-color screen. It is rather surprising, therefore, that no one 
in the generation since Ives disclosed this idea, made any attempt to utilize it. 
J. E. Thornton (Eng. P. 238688) proposed to use a screen plate for two colors 
in a three-color process, but the third color was to be obtained by a separate 
beam. There is very little economy in this type of procedure. The same in- 
ventor also proposed upon many occasions to use a double-width film, each half 
of which contained a two-color line screen. A beam-splitting camera projected 
an image upon each half of the film, giving the possibility of a four-color 
process. 

It was not until the Ives patent had almost expired, that the original dis- 
closure became resurrected. G. S. Whitfield, of the Ilford company (Eng. P. 
371009) proposed to use a bipack loaded celluloid to celluloid. The front ele- 
ment was blue-sensitive. The rear element was red- and green-sensitive, and 



146 HISTORY OF COLOR PHOTOGRAPHY 

carried a screen between the base and the emulsion. This scheme is different 
from the original, and also not nearly as effective. The existence of two thick- 
nesses of celluloid between the front emulsion and the rear certainly does not 
tend to improve definition, already rendered somewhat inferior by diffusion 
through a front emulsion. Neither is complete registry possible, since this 
separation allows a further spreading of the rays, making the rear images 
slightly larger than the front. 

A method of printing from this bipack is disclosed by Mr. Whitfield in 
English patent 416114. Another variation of the Ives original disclosure is 
contained in English patent 419887, issued to J E. Thornton. He proposed to 
put a red-sensitive emulsion in front of a two-color screen which registers the 
other two primaries. As is the case with the Whitfield disclosure, this one is 
just so much wishful thinking. All the objections mentioned previously hold 
true here, and to these can be added the most serious one of all, that of poor 
color separation. It is not possible at this writing (1943), to prepare an emul- 
sion in which the red sensitivity is so much greater than the blue that no 
interference will result. It has been proposed (cf. chapter on Tripacks and 
Bipacks) to dye a fast red-blue-sensitive emulsion, with a yellow dye in order 
to reduce the blue sensitivity. That most certainly can be done, but then very 
little blue light will reach the emulsion below. The loss of speed would be so 
great that even the most inefficient of one-shot cameras would be fast in com- 
parison. The proposed improvement upon the Ives screened bipack is an 
improvement in name only, and represents merely someone's idea that it 
would be nice if the thing worked. But it doesn't, not anywhere near as well 
as the original. 



CHAPTER 13 

FORMATION OF THE SCREEN 



I. 



LN the previous chapter there were outlined eight general classes into which 
the techniques used for the making of the screen could be grouped. Here will 
be discussed the specific methods contained in each of the groups. The first to 
be discussed will be those procedures that do not yield geometrically patterned 
screens. Of these the most important are the ones utilizing "dusting-on" 
schemes. J. W. McDonough, who shares with Prof. Joly the distinction of 
being the first to put du Hauron's suggestions into practice, was the first to 
use this technique (Eng. P. 5597/92; U.S.P. 411186 and 471187). The surface 
of a plate or film was made tacky by a coat of varnish. On this was dusted a 
mixture of fine colored powders made of glass, pigments, stained gelatin, resins, 
or shellac. The last was the preferred medium. Three batches of shellac were 
dissolved in alcohol, dyed primary red, green, and blue respectively, dried, and 
finally ground into fine powders. These were mixed in the proportions required 
to fulfill the first black condition, then applied to the tacky surface. If any 
colorless interstices showed up, heat was applied up to the point at which the 
colored powders fused together. He also suggested that the emulsion surface 
could be rendered tacky, and the colored particles dusted over it. This would 
allow the exposures to be made through the surface of the emulsion, in the 
normal manner. But if the colored particles consisted of shellac or other water- 
proof material, some difficulty might be met in the processing. Presumedly, 
in this case, it is the stained gelatin powders that were to be used. 

Soon after 1900 came the Autochrome plate which was the discovery and 
development of the Lumieres, assisted by Seyewetz. This plate enjoyed a 
considerable popularity from the very beginning, and its use extended well 
into the nineteen thirties. The elements used were starch grains (Eng. P. 
22988/04; U.S.P. 822532; Fr. P. 339223; Ger. P. 172851), ferments, yeast, 
bacilli, or powdered enamels. The grains were dyed magenta, cyan, and 
yellow, mixed to yield neutrality, and then spread over a tacky surface. On 
top of this was coated a second tacky surface upon which was distributed dyed 
grains such that the two together would yield red, blue, and green elements. 
The refractive index of the varnishes must be the same as that of the color 
elements, otherwise a loss of light by interface reflections would result. The 
same result could be obtained by mixing particles dyed in six colors, red, 
orange, yellow, green, blue and violet. (Eng. P. 25718/04, 9100/06 and 
8153/11; U.S.P. 877351). 

147 



148 HISTORY OF COLOR PHOTOGRAPHY 

Some interesting details concerning the preparation of the plates were given 
by E. Ventujol (Rev.franq. Phot., Vol. 4 (1923), p. 80; E. J. Wall, "History of 
Three-Color Photography" (1925), p. 467). The grains were sifted by elutri- 
ation, and dyed with triphenylmethane dyes. The spaces between the colored 
particles were filled in with charcoal, which was applied by means of a special 
machine. A daily examination of the plates was made to check the procedure, 
and one day the grains appeared to be much more transparent than usual. 
This was traced to the application of pressure to the plate, a procedure that 
was finally adopted as routine. About 5000 kilos (approximately 1 1 ,000 pounds) 
pressure per square centimeter was applied. The emulsion was one rich in 
silver bromide, and poor in gelatin. 

The Lumieres were not the only ones to use yeast or other dead cells. Borrel 
and Pinoy (Fr. P. 373492) stained microbes, or staphylococci which had 
previously been killed by heat. In French patent 379632 they also suggested 
yeast cells. A variation of this was disclosed by H. Clement (Fr. P. 416700). 
He allowed the corpuscles to grow to the proper size, dyed them while still 
alive, then killed them. Dyed pollen was suggested by K. Campbell (Eng. P. 
224571). The pollen was bleached in peroxide, dyed, mixed to yield a neutral 
black, dispersed in a colloid, then spread upon a glass or film surface. Ten 
years later (U.S.P. 1965852) M. 0. Miller patented practically the same pro- 
cedure. The use of pollen, starch, microbes, etc., made unnecessary the 
grinding operation for the preparation of fine powders. 

Many experimenters used gelatin as the color-carrying medium. McDon- 
ough, as was shown above, appears to have been the first person to suggest its 
use. Some years later J. Bamber (Ger. P. 233140; Eng. P. 3252/08 and 
rii47/o9; Fr. P. 399320) disclosed a more specific procedure. Dyed gelatin 
was tanned by treatment with formaldehyde, ground to a fine powder, and then 
mixed to yield a neutral black. This was finally sifted on to a celluloid surface 
that had previously been made tacky by a coating of gum. In a later patent, 
however, (Eng. P. 15775/n) he suggested the replacement of the gelatin by 
gum sandarac. J. Szczepanik (Ger. P. 354317; Eng. P. 17065/08; U.S.P. 
1089602) varied the procedure somewhat. Three batches of gelatin were dyed 
with basic dyes, then ground to powders. When the dyed gelatin was dusted 
over a collodion surface and came into intimate contact with that surface, the 
basic dye transferred from the gelatin to the collodion. When sufficient dye was 
thus transferred, the gelatin powder could be removed, leaving a dyed-screen 
pattern upon the collodion. 

Powdered dichromate-hardened fish glue, dyed presumedly with a pinatype 
dye, was used by F. M. Duncan (Eng. P. 50/09) for the preparation of the 
screen elements. H. W. H. Palmer (Ger. P. 233167; Eng. P. 6279/11; Fr. 
P. 406147) tanned gelatin, gum arabic, tragacanth, etc., with formaldehyde. 
These were dyed, then sprayed into a drying chamber. In this way there were 
obtained very fine powders. Or liquid glass, properly colored, could be blown 
into fine powders, then mixed and spread over a tacky surface. In order to 



FORMATION OF THE SCREEN 149 

fill in the interstices, heat was applied to fuse the elements together. This had 
been previously suggested by E. Gistle (Ger. P. 228597) who also added a 
final touch of grinding the glass elements smooth before coating them with an 
emulsion. 

A slight novelty was introduced by H. N. Holmes and D. H. Cameron (U.S.P. 
1429430). The dyed gelatin particles were dispersed in gum dammar, and the 
colloidal solution coated in a single-grain layer upon the plate or film base. 

Atomization is evidently a very convenient method for the formation of 
powders, since so many inventors seem to turn to it. C. Rusicka (Eng.,P. 
326764, 326779, 326780, and 326781) atomized dyed gelatins to form the 
powders. These were mixed to yield neutrality, then dispersed in nitrocellulose 
and coated upon the emulsion side of a panchromatic plate. This creates a 
serious problem of processing, for the screen elements would form a waterproof 
coating over the emulsion. W. R. B. Larsen (U.S.P. 1918208) would atomize 
dyed gelatin to which the proper optical sensitizers had been added. The dried 
gelatin powders were then mixed and spread on a film base. Over this would 
be coated a color-blind emulsion. The liquid gelatin would dissolve the dye 
plus sensitizer, which would then diffuse into the emulsion. Each emulsion 
grain would be sensitized and colored properly. One drawback of a scheme 
such as this is that only a few layers of emulsion grains would be sensitized and 
dyed, thus seriously reducing the already limited latitude of the plate. 

Instead of tanned gelatin, H. W. H. Palmer (Eng. P. 22228/07) proposed 
to use stained fluxes such as fused borax beads, or ceramics. These could be 
sifted on to a glass surface, then fired until fusion took place. A similar pro- 
cedure was suggested by C. F. Bleeker (U.S.P. 11 75224). He made colored 
powders from glass, mica, fused borax, etc. Zinc sulphide grains, upon which 
had been adsorbed aluminum hydroxide (a strong absorbent of basic dyes), 
was suggested by L. Paris and G. Piard (Fr. P. 477173, addition 20019). A 
similar idea, except that calcined magnesia formed the dye-absorbing grains, 
was disclosed by B. T. Silverman (Ger. P. 313008). Somewhat along these 
lines was the scheme proposed by A. A. Canton (U.S.P. 1828958). Iron parti- 
cles were oxidized, then stained and spread on a surface receptive to the dye. 
The color transferred from the iron particles to the surface, after which the 
iron was removed by a magnet. This last is a variation of a much earlier 
proposal by Szczepanik which was described above. M. Wieland (Eng. P. 
I 375° 2 ; Ger. P. 353759; U.S.P. 1497747) proposed to use the synthetic 
resins made by the condensation of phenol with formaldehyde. 

A rather novel scheme was suggested by J. Camiller and A. Hay. This was 
especially suitable for two-color processes, but could be applied to three colors 
as well. Particles of one color (Eng. P. 154150 and 158670; U.S.P. 1440373) 
were dispersed in a colloid dyed another color. The particles must not interact 
with either the dispersing medium or the dye it contained, otherwise faulty re- 
sults would be obtained. The mixture was then coated on to a support in the 
form of a single-layered emulsion. It is rather difficult to see how to prevent 



ISO HISTORY OF COLOR PHOTOGRAPHY 

each particle from becoming completely surrounded by the colored dispersing 
medium. C. Schleussner (Ger. P. 293004) sifted the colored elements over a 
surface coated with acetic acid and glycerin. The acetic acid was then allowed 
to evaporate. The application of steam caused the particles to swell, thereby 
closing up any interstices left between the grains. 

Celluloid and the cellulose esters in general, could also be used. F. L. Dyer 
(U.S.P. 947965) atomized a solution of celluloid in amyl acetate, forming 
transparent colored particles that were very finely powdered. A similar scheme 
was outlined by I. Kitsee (U.S.P. 1206000). Solutions of dyed celluloid or 
gelatin were sprayed into a high compartment. Very fine particles were col- 
lected at the base. These were mixed and sprayed to a tacky surface. In an- 
other disclosure (U.S.P. 1 221457) he proposed to coat paper with gum, then 
with colored celluloid. This was cut, after drying, into strips 0.00125 to 0.002 
inches square. The pieces were soaked in water, which dissolved out the gum, 
then mixed and applied to a support. R. Lehner (Eng. P. 7629/08) ground 
dyes into celluloid by grinding with alcohol-water solutions. After homogeneity 
was achieved, the water was driven off, and the celluloid worked into a film. 
S. E. Sheppard, of the Eastman Kodak Company (U.S.P. 1290794) dyed nitro- 
cellulose with basic dyes, and emulsified the result in a colloid. The dyes used 
were the Sudan colors which were alcohol-soluble. Water-soluble dyes could 
be made spirit-soluble by precipitation with tannic acid. The action of this 
substance upon gelatin was prevented by a further treatment with antimony 
salts. The dyed nitrocellulose could be dispersed in gelatin dissolved in 
water, in which nitrocellulose was completely insoluble. C. L. A. Brasseur 
(U.S.P. 974464; Eng. P. 18750/08) formed threads out of colored celluloid. 
These were rolled on to a sheet of paper with a tacky surface, and transferred 
from this to a celluloid film. 

A substance eminently fitted for use as a carrier for the color elements is 
resin, and its related compounds. Like nitrocellulose, it is a waterproofing 
material so that the colors held within it would be insulated from the action 
of the aqueous processing solutions. It can also serve well as a substratum for 
the emulsion, so that it can replace the usual subbing. One of the first to sug- 
gest this substance was McDonough (cf. above). Christensen appeared to 
have concentrated much of his effort along these lines. We will discuss his 
disclosures a little later, since they were adopted by the Agfa company in the 
production of its screen plate. This product was quite popular until the 
advent of the new monopacks, which, overnight, made all other color- trans- 
parency processes practically obsolete. 

H. W. H. Palmer (Eng. P. 16313/09), two years after he proposed to use 
tanned gelatins, turned to dyed caseins and the gums, of which resin is a 
specific example. These were dyed, made insoluble, pulverized, mixed, and 
finally spread upon the tacky surface. Upon standing, the elements absorbed 
water, swelled, and in that manner closed any clear spaces left between them. 
Pressure applied to the screen elements, flattened them out evenly. A. Wieb- 



FORMATION OF THE SCREEN 151 

king (Ger. P. 250036) formed fine powders by spraying the colored resins into 
a liquid that would coagulate them. This was later adopted as a good working 
rule by S. E. Sheppard, who used nitrocellulose instead of resin, and who 
caused the precipitation to take place inside of a colloid such as gelatin, in 
which the nitrocellulose was insoluble, and with which it was immiscible. 
This was an improvement on the general Wiebking procedure. The Pathe 
Company also used resins to carry the color elements (Ger. P. 261 161). K. Froh- 
lich emulsified dyes in solutions of resins or oils. These were sprayed on to a 
support, and the dyes transferred themselves to the support through the emulsi- 
fying substance, which could be removed after the transfer was complete 
(Ger. P. 406706). 

In 1908 J. H. Christensen proposed to dye dextrin and gum arabic, and 
emulsify the dyed substances in gum dammar which was thinned with turpen- 
tine. (Eng. P. 20971/08; Ger. P. 224465; U.S.P. 994977). The particles were 
allowed to settle, after which they were suspended in benzene, carbon tetra- 
chloride, or kerosine and asphalt. A mixture of the particles was poured over 
a support coated with rubber. The dyed globules could be tanned (Eng. P. 
21007/08 and 7480/12) with substances that would have no action upon the 
material of the support. In some later disclosures (Eng. P. 1633 n and 216853; 
U.S.P. 1486635; Ger. P. 403590) he proposed to form dye tannates, etc., in 
alcoholic media. The solutions were then atomized or emulsified, and the 
particles applied to a support. Again, in 1925 (U.S.P. 1564202) he returned to 
the problem of making fine powders. This time he emulsified a liter of a dye 
solution in three liters of thick rubber solution, or rubber-resin mixture. The 
rubber was removed by carbon tetrachloride, leaving a finely-ground colored 
powder behind. After mixing properly, the powders were applied to the car- 
rier. In 1929 (Eng. P. 3091 13) he prepared the screen elements by still a 
different method. Basic dyes and tannic acid were dissolved in alcohol, then 
emulsified in gasoline. The emulsions, properly mixed, were sprayed on to a 
glass surface. Upon drying, the dye and the tannic acid formed a transparent 
pigment. It was also possible to use acid dyes together with the basic. Suit- 
able dyes were brilliant crocein, patent blue, sea green, erio-glaucine, and 
naphthalene green. 

The Agfa company proposed (Ger. P. 254180; Eng. P. 9167/12) to atomize 
colored dextrin solutions, by means of compressed air saturated with water. 
The particles fell on a surface that was inclined, and over which flowed a layer 
of turpentine. This company controlled the Christensen screen-plate patents, 
and manufactured a plate which was quite successful up until the late nineteen- 
thirties. 

A method which suggests itself immediately for the formation of a screen, 
is the use of colored transparent fabrics woven from threads. The silk industry 
has achieved a high degree of skill in the preparation of very fine fibers, and 
if this be composed of the artificial silks, they would be very transparent. It 
was Prof .'J. Joly, one of the first independent workers to put du Hauron's ideas 



152 HISTORY OF COLOR PHOTOGRAPHY 

into practice, who suggested the uses of silk fibers for the screen elements 
(Eng. P. 19388/95). His first suggestion was to apply the colored threads to 
a tacky surface, a sort of dusting-on procedure. But he also suggested that 
the fibers could be woven into a cloth, all the colored fibers running in one 
direction, while the fibers running across these were to be colorless and trans- 
parent. R. Berthon and J. Gambs (Eng. P. 20834/06; Fr. 357928, additions 
5986 and 9464) also wove fabrics out of colored threads, making sure that the 
colored fibers lay adjacent to each other so that no clear spaces would exist 
between them. 

Instead of weaving the fibers into a fabric, Ratignier and Pervilhac placed 
them in celluloid just as that was being cast into a film. The comb, through 
whose teeth the fibers passed, was so arranged that the fibers touched each 
other at the point where they entered the celluloid solution. The fibers were 
thus kept in position by the solidified celluloid (Fr. P. 391785). A somewhat 
similar idea was disclosed by E. B. Smith (Eng. P. 139871). In this case a glass 
plate with a layer of gelatin was used instead of celluloid, and the comb was 
arranged so that the fibers touched each other just at the point where they 
came in contact with the gelatinized base. 

F. Fritz, in a general discussion of the problem (Ger. P. 227130) pointed out 
that a woven fabric had the disadvantage of having a large number of clear 
areas not covered by colored fibers. He overcame this difficulty by the use 
of the artificial fibers. After the cloth was woven, he applied heat which 
softened and spread out the individual threads and caused fusion to take place. 
He also proposed to weave one thread of a transparent material and then 
imbed the cloth in a plastic dyed with the missing color (Ger. P. 231676). 

Instead of using textile fibers for the weave, A. J. Jorelle (Fr. P. 460724) and 
A. N. Pierman (U.S.P. 1196718) suggested spun glass. The French patent 
appeared in 191 2, while the United States patent did not appear until 1916, 
but it was applied for in 1908. Therefore the two inventors worked in com- 
plete ignorance of each other's contribution. Despite the apparent simplicity 
of this procedure, no successful plate made this way, was ever marketed. 

The next method to be discussed includes the processes in which a block is 
formed by cementing together layer upon layer of dyed gelatin or collodion 
sheets. This was apparently first suggested by H. Snowden Ward {Brit. J. 
PhoL, Vol. 46 (1899), p. 399). He proposed to stain thin gelatin sheets in the 
three primary colors, then superimpose them so that the colors alternated 
until a block was formed whose depth was at least equal to the length of the 
projected screen plate. If each sheet is to be 0.001 inch thick, it would require 
1000 sheets to make a block one inch deep. If the plate was to be seven inches 
long, then about 7000 sheets would be required. Technological progress in 
casting cellophane, celluloid, etc., has been so great that the making of large 
sheets that have a maximum thickness of one thousandth of an inch is common- 
place. The gelatin emulsion layer, coated on positive motion-picture film 
stock, is only a half of a thousandth of an inch thick, and the individual layers 



FORMATION OF THE SCREEN 153 

used in a Kodachrome or Ansco Color monopack are even thinner. It is not a 
difficult matter to form such blocks. The writer remembers a summer which 
he spent in the sheeting department of the Celluloid Company of America 
where it was a commonplace affair to cut sheets of celluloid, 36 by 72 inches, 
each 0.001 inch thick, and superimpose them to form a new block of celluloid 
that contained a complicated design. The new block could then be sliced as 
desired. In order to form sections suitable for screen plates, the block must 
be cut in a direction perpendicular to the plane of the colored sheets. 

At a later date {Phot. /., Vol. 40 (1900), p. 141) Ward attributed the idea to 
Prof. Joly. However, sometime before this, O. H. Witt had already made a 
German patent application (Ger. P. Anm. 14,564 IV, 57a) for an idea identical 
to this, using stained celluloid or gelatin sheets for the elements. The applica- 
tion was abandoned on May 8, 1899. Some time later R. E. Liesegang (Jahr- 
buch, Vol. 22 (1908), p. 147; Phot. Almanach, 1904, p. 122) proposed the same 
idea. However he suggested that each layer might be coated on top of the 
other, thus insuring good adhesion and the absence of clear spaces. The idea 
was again suggested by Krayn, who marketed plates for a short period of time 
(Ger. P. 167232; Eng. P. 19202/05; Fr. P. 357895, additions 5375, 6534, 6536, 
6537; 386772 and 409367). Or the colored sheets could be cast separately 
(Ger. P. 167613) and then assembled to form a block that would be cylindrical 
in shape (Eng. P. 495/07; Ger. P. 18843 1, I 9°56o; Fr. P. 388913). By this 
procedure long film strips could be cut. If a block were formed, and this cut 
horizontally, then reassembled into a new block whose dimensions were ex- 
actly opposite to the first, the new block could be cut to give a screen consisting 
of a definite mosaic pattern, rather than a straight-line pattern (Ger. P. 193463 ; 
Fr. P. 405924; Eng. P. 2213/08). The Krayn disclosures were put into com- 
mercial use by the Neue Photographische Gesellschaft, who later attempted 
to put into practical use the Fischer and Siegrist color development ideas (cf . 
chapters on Color Development). 

The same scheme of reforming the block and cutting again but in a different 
plane, was proposed once more by F. A. Fifield, (U.S.P. 990247) and somewhat 
later, by C. L. A. Brasseur (Eng. P. 28631/13; U.S.P. 1081484; Fr. P. 466120). 
The last proposed to make some of his sheets 0.00025 inch thick. Since it 
was impossible to cut the sheets so thin, he sliced them as thin as he was able, 
then polished them to the desired dimension. This made the process rather 
difficult and expensive. It is rather doubtful even with all our later ability to 
make fast fine-grained emulsions, whether elements as fine as 0.00025 inch 
could be used. Assuming that the average grain has a diameter of 0.5 /x, there 
would be approximately 12 grains of emulsion across each dimension of the 
element, or altogether 144 grains. This is by far too few to represent a good 
average of the emulsion, and also much too few to yield an appreciable density 
behind the element. This number of grains also fails to take into considera- 
tion the fact that the grains do not lie end to end, but are separated by a wall 
of gelatin whose dimension is at least equal to that of the grain itself. There- 



154 HISTORY OF COLOR PHOTOGRAPHY 

fore it would be closer to the truth to state that in the area behind each screen 
element, which is 0.00025 inch square, there would be at most fifty emulsion 
grains. If the speed of the plate were of no consequence, much finer grains 
could be used, such as those present in the Lippmann colloidal emulsion. Then 
screens of much finer texture could be used. 

W. C. Masser and W. Hudson (Eng. P. 25730/07) also adopted the Krayn 
idea of making long films by cutting the blocks circumferentially. These two 
worked independently of Krayn and were obviously ignorant of his disclosures, 
since both patents were disclosed to the public in the same year. The Ver- 
einigte Kunstseidefabriken cast color sheets, presumedly of cellulose acetate or 
artificial silk, that were approximately 0.13 mm in thickness, or 0.005 inch. 
These were superimposed before they were completely dry. 

Perhaps this idea should really be classified under the miscellaneous pro- 
cedures, but since the final screen is formed by cutting sections from a block, 
it is included here. It was disclosed by F. Fritz (Ger. P. 218324 and 223819) 
in 1908. Artificial threads, such as those made from cellulose acetate, rayon, 
silk, etc., are woven into sheets. Or they can first be woven into strands, then 
combined to form a solid block that is finally cut into sheets by means of a 
microtome. The weaving technique enabled the formation of any desired 
mosaic pattern. 

A variation of the procedures outlined above is contained in the disclosure 
of S. Schapovaloff (Eng. P. 389265). He placed between the colored sheets 
a sheet that had been dyed black, so that the order of the colors became red, 
black, green, black, blue, black, red, black, etc. The block was cut spirally to 
obtain long rolls of film. The inclusion of black between the elements was to 
insure against distortion due to a spread of the light beam on its passage 
through a color element. But if the colors used were mutually exclusive, as 
they should be for best results, a green color-element would act just like black 
for the light that passes through either the red or the blue element. 

The use of glass powders for the formation of the screen, has long been 
known. A. G. Oneil (U.S.P. 1892901) would cover an area with colored glass 
hexagonal rods, fuse them together by heat, then draw out the entire bundle 
of rods to the proper size, and slice off sections. 

The ability of the photographic emulsion, be it sensitized either with di- 
chromated gelatin or silver halide salts, to reproduce fine details, was quickly 
seized upon for the formation of screens. It was first suggested by du Hauron 
("Les Couleurs en Photographie Solution du Problem/' 1869, p. 54; "La 
Triplice photographique et lTmprimerie," (1897), p. 336). A gelatin layer, dyed 
primary red, and Sensitized with dichromate, was exposed through a black- 
and-white line screen, the blacks having twice the width of the whites. The 
action of the light tanned the gelatin that corresponded to the transparent 
areas of the master screen. The soft-gelatin was removed by a hot water etch. 
This left a series of red lines, separated by transparent portions whose width 
was twice that of the reds. The plate was next recoated with gelatin dyed 



FORMATION OF THE SCREEN 155 

green, and sensitized with dichromate. The second exposure was also through 
a black-and-white screen, but this time the width of the two was the same, 
and their direction was at right angles to the red lines. Since the red portions 
passed no light that was effective upon the dichromated gelatin, only the 
gelatin lying between the previously formed lines was affected, and of this, 
only one half of the total area became tanned. Hot water treatment removed 
the unaffected gelatin, which was finally replaced with new gelatin dyed a 
primary blue, and as before, sensitized with dichromate. The last exposure 
was made without any master screen, only through the red and green areas 
formed previously. These acted as opaque lines for the blue rays to which the 
gelatin was sensitive. Between all coatings, it was desirable to further treat 
the plate to thoroughly fix the dyes or coloring matter in the tanned gelatin. 
Hence no diffusion of colors was possible. 

Although du Hauron completely outlined the methods, he did not put them 
into practice. E. J. Wall attributes to C. L. A. Brasseur the honor of being 
the first person to use dichromate photography for the formation of the screen 
("History of Three-Color Photography," 1925, p. 483). The du Hauron tech- 
nique was slightly changed in that after each printing a layer of varnish was 
put upon the plate to insulate one color element from the others (Eng. P. 
21210/09; Ger. P. 179378). This was in 1904, a considerable time after the 
first du Hauron disclosures. E. Sanger-Shepherd (Eng. P. 20384/07) did not 
depart in the least from the du Hauron specifications. A prolific patentee in 
this field was J. H. Powrie. Instead of gelatin, he preferred to use fish glue 
(U.S.P. 802471; Ger. P. 215072, 222504 and 225004; Fr. P. 358746, 358747; 
Eng. P. 20662/05). The replacement of the gelatin by the fish glue allowed 
him to develop in cold water. The insoluble glue was then dyed up with an 
acid dye such as aurophenin. This was then treated with a basic dye, either 
brilliant green, safranin or thiazine red. A reaction between the acid and basic 
dye took place which yielded an insoluble molecular combination of the two. 
Had Powrie used dyes of the same hue, an intensification of the color would 
have resulted. But in this case, the yellow served to convert the red or the 
*green into shades that matched the primary colors. The dyed lines were further 
pigmentized by treatment with tannic acid, after which the plate was resensi- 
tized and exposed under the screen again, and the entire procedure repeated 
until three sets of lines were obtained. 

In some later patents Mr. Powrie discloses a novel scheme for printing the 
screen lines in juxtaposition (U.S.P. 1605062, 1717404; Eng. P. 287188). A 
celluloid base was coated with a light-sensitive emulsion on one side, and on 
this was printed a master matrix in which the transparent lines had one half 
the width of the opaque lines. On the other side of the celluloid film base was 
coated a dichromated gelatin. The first printing was made with a beam of 
light whose axis was perpendicular to the plane of the celluloid base. Therefore 
there was impressed on the other side a series of lines consisting of tanned 
gelatin which lay directly below the transparent lines of the master matrix. 



156 HISTORY OF COLOR PHOTOGRAPHY 

These lines were dyed with a primary color and were given an added stability 
by treatment with a tannic acid mordant. This was recoated with dichro- 
mated gelatin, and exposed again through the same matrix, but now, instead of 
using a beam of light whose axis was perpendicular to the film base, a beam 
was used whose axis was at an angle such that the zone of action of the light on 
the dichromated gelatin was adjacent to the line already formed. By this 
means, a line of tanned gelatin was formed next to the first one. After removal 
of the unaffected gelatin the new set of lines was dyed in a second primary 
color, thoroughly tanned, and finally recoated with dichromated gelatin. The 
third printing was done after the master matrix had been buffed away, through 
the two sets of lines already formed. These were dyed red and green respec- 
tively, so that as far as their action on the gelatin was concerned, they were 
completely opaque. Screens with upward of 1200 lines to the inch have been 
made by Mr. Powrie using this technique. 

A really novel scheme was proposed by C. E. K. Mees (Eng. P. 28406/07) 
and assigned to the Wratten & Wainwright company. A single exposure 
sufficed to give three differentially hardened lines. This was accomplished 
by using a screen that was composed of an opaque, a semi-opaque, and a 
transparent line. Dichromated gelatin was then exposed behind this master, 
and after washing to remove unaffected dichromate, dyed up with patent 
blue. A thorough wash left the following condition. The completely tanned 
portion of the gelatin was dyed a deep blue. The partially tanned line absorbed 
considerably less dye, and this appeared as a blue-green, or cyan. The soft 
line (completely un tanned) remained undyed. Upon soaking in a yellow pina- 
type dye, the medium and soft portions turned green and yellow respectively, 
while the hard line was unaffected. A final dyeing with a red dye, which would 
not take in tanned gelatin at all, completed the screen. This is really an in- 
genious scheme, and the only drawback would appear to be the fact that 
patent blue would always have some green transmission, and thus an impure 
blue primary rendition would result. 

F. Faupel (Ger. P. 216610 and 220154) also took advantage of the differential 
dyeing properties of soft and hard gelatin. Dichromated gelatin was exposed 
under a line screen and then washed to remove the unaffected salt. This gave 
a series of hard lines separated from each other by lines formed of soft gelatin. 
The tanned gelatin lines were then dyed with azorubin or crystal ponceau 6R, 
while the soft gelatin lines were dyed with methyl blue, brilliant wool blue G, 
or brilliant azurine. After dyeing the two sets of lines, the plate was resensitized 
and exposed under a line screen, but at an angle to the first, then dyed with fast 
chrome yellow G. This converted the untanned portion of the lines into a 
green, which would indicate that the blue used was really a blue-green and not 
a primary blue. The screen was composed of red, green, and cyan lines, a poor 
combination for color. The exposed and washed plate could also be dyed 
with a mixture of naphthol green and crystal ponceau. These dyes stained 
either tanned or untanned gelatin only, so that the tanned lines were colored 



FORMATION OF THE SCREEN 157 

red, and the untanned were colored green, simultaneously. Resensitization 
with dichromate was followed by exposure to a master, but at an angle to the 
first. This tanned part of the green line. The naphthol green of the untanned 
portions was removed by treatment with alcohol and water, after which it was 
dyed with brilliant azurine. This had no action on the tanned portions of the 
plate, dyeing up only the clear parts. 

The same procedure was outlined again by C. Spath (Eng. P. 3601/09; U.S.P. 
946470 and 1292347). Mr. Spath even fell into the same error of converting 
part of the blue line into a green by treatment with a yellow dye, indicating 
that he also used a blue-green rather than a primary blue for one of the ele- 
ments. 

These procedures were slight variations of the one disclosed by Berthon and 
Gambs (Fr. P. 358250) several years prior. Dichromated gelatin or other 
colloid, dyed "red," was exposed under a line screen. After the exposure, the 
dye was washed out of the unaffected regions, which were then dyed "blue." 
A protective varnish coating was put on this, and another layer of dichromated 
gelatin coated on top. This time a yellow dye was present. The exposure was 
made with the lines at a forty-five degree angle to the first. The untanned por- 
tions of the second layer were dyed "red." These colors were not really red 
and blue, since the "red" by the superposition of yellow, became primary red, 
and the "blue" became green. Hence the "red" was really magenta and the 
"blue," cyan. Where the "red" and "blue" overlapped, there was formed 
blue-violet or primary blue. This was another indication that magenta and 
cyan were the true shades. 

Only slightly different from this is the procedure outlined by J. M. Borrel 
(Fr. P. 393557). Dichromated gelatin was exposed under a screen. The tanned 
portions were stained with a red dye while the untanned portions were stained 
"blue." Resensitization followed, and this time the portions of the blue lines 
that remained untanned were stained with a yellow, which converted the 
" blue " into green. The last step was to treat the plate with a " red " dye which 
would affect only the hard gelatin areas. Therefore the green areas were un- 
affected. Staining a red area with a "red" (really magenta) dye left its shade 
unchanged. But the "blue" or cyan areas were converted into primary blue. 
The loose usage of the terms "red," and "blue" leads to considerable confusion, 
and care must be taken to realize when a primary red or when a magenta or 
secondary red is really meant. B. Bichtler (Ger. P. 292347) used a celluloid 
base coated on both sides with dichromated gelatin, and formed one series of 
color lines on each side. The third series was put on after one side of the film 
was recoated with dichromated gelatin. 

A plate that enjoyed a large measure of success until Dufaycolor was intro- 
duced, was produced under the Finlay patents. The earlier patents utilized 
the dichromated gelatin technique, although later disclosures indicated that a 
resist method was adopted. This last will be discussed below. The earlier 
product was novel in that a mosaic of red and green circles upon a blue back- 



158 HISTORY OF COLOR PHOTOGRAPHY 

ground was used. The screen was mounted upon a glass plate, and used as a 
front element in a bipack, the rear element of which was a panchromate emul- 
sion (Eng. P. 19652/06, 4208/10; Fr. P. 426498; U.S.P. 1085727). 

I. Kitsee (U.S.P. 1383819, 1426995, and 1426996) took advantage of the fact 
that tanned gelatin would not swell. The soft lines did swell, so that they 
stood out in relief. These were dyed by means of a roller. After dyeing the 
swollen portions, the spaces between were filled with another color. A two- 
color screen is described by J. G. Zimmermann (U.S.P. 1579464). A prolific 
patentee is J. E. Thornton, who seemed to specialize in making screens for a 
four color process (U.S.P. 1670671 and 1873673; Eng. P. 224571, 253643 and 
257836). He favored the use of fish glue rather than gelatin. After exposure, 
the dye was washed out of the soft areas, and they were then redyed with a 
pinatype dye. The film base was double width, and only half as thick as the 
normal. Each half contained two of the four colors. After processing, the 
film is slit in two and cemented together, either face to face, or celluloid to 
celluloid. 

In a patent issued to the Durkoppwerke, A. G. (Eng. P. 389843) it was 
suggested to have one half of the screen area undyed. To this end, a film of 
dichromated gelatin was exposed to a master. The untanned portion was 
dyed, then resensitized and exposed. The dye was washed out of the un- 
affected parts, which were then redyed in a complementary color. The use of 
such a screen would lead to poor color separations since both colors would be 
rendered behind the transparent line, indiscriminately. This procedure was 
also applied to £he making of paper prints (Eng. P. 389844). 

Instead of using dichromated gelatin for the light-sensitive medium in the 
preparation of the color elements of the screen, it is possible to use silver-halide 
salts. The first to suggest this appears to have been L. Gimpel (Fr. P. 414953). 
A screen of parallel lines is photographed upon a fine-grained silver-halide 
emulsion. The silver image is then toned a primary red. After recoating, the 
screen was photographed again, but this time at right angles. The silver was 
then converted into a cyan color. After recoating the plate a second time, 
presumedly with an emulsion sensitive to the blue only, the plate was exposed 
through the lines already formed, so that behind the red line there would be 
formed no deposit, behind the cyan line a "half tone" deposit, and in the 
remaining area, a full tone. This was converted into a yellow.. Although the 
process appears hopeful, the toning seems to be such that a very poor screen 
results. Two of the elements are approximately correct. The initial exposure 
yielded a primary red, while the second exposure yielded a cyan which was 
converted into a green by the third treatment. But the third area in the unit 
will be yellow, whereas a blue color is demanded here. 

A combination of dusting-on and chemical toning processes was adopted by 
E. H. Tarlton (Eng. P. 1 10993). A transparent support was coated with a 
layer that contained red particles dispersed in it, so arranged that they blocked 
out only half the surface area of the film. On top of this was coated a red- 






FORMATION OF THE SCREEN 159 

insensitive silver-halide emulsion. The exposure was made through the red 
elements, so that silver images were formed that completely filled in the spaces 
between the red particles. This was converted into an iron blue. Mr. Tarlton 
proposed his scheme for two-color analysis, hence the use of red and minus-red 
elements was correct. G. Valensi (Fr. P. 536737) also used a base coated on 
both sides with silver-halide emulsions. A line screen was photographed on 
each side, and the silver on one side converted into a yellow color, while on the 
other side it was converted into a "red." After resensitization, another series 
of lines were printed on the film, and toned "blue." Here again, a false color 
analysis would be obtained, since yellow forms one of the screen elements. 

A silver-halide emulsion again forms the basis for screen formation in a dis- 
closure issued to Naturfarben-Film G.m.b.H. The silver comprising the screen 
elements is treated with a tanning bleach. The soft gelatin is washed away, 
and the relief images remaining are dyed (Ger. P. 382575). This procedure is 
also utilized by H. Butschowitz (Ger. P. 389852), but instead of using a tanning 
bleach, he used a tanning developer. The chemical principles governing such 
processes will be discussed in a later chapter. 

L. F. Douglass (U.S.P. 1504465) printed a dot screen on to a silver-halide 
emulsion, and dye-toned the resultant silver image with a magenta dye. The 
clear areas between these were sensitized with ferric salts, and after exposure 
developed with ferricyanide, forming the well-known iron blue. The entire 
film was finally dyed yellow, which converted the iron blue into a green, and the 
magenta into a red. H. Muller (U.S.P. 2007282) exposed a silver-halide 
emulsion and obtained a screen composed of silver elements. These were toned 
one color, and the clear areas between them, dyed another. It was also possible 
to use a tanning bleach, such as is used in the ozobrome processes, to convert 
the gelatin in situ with the silver to a tanned form. This can be dyed with a 
color which is retained by tanned gelatin, but which is washed out of the soft 
areas. These last can be treated with pinatype dyes. This is merely another 
form of the technique disclosed by Mees. 

It is not a far cry from photographic to photomechanical printing methods, 
which by 1900 had been developed to a high state of mechanical perfection. 
Not many of the patent disclosures deal with such schemes, principally because 
the printing of more than 175 elements to the inch creates many problems whose 
solutions had not been solved up to forty years later. The one advantage of 
this procedure would be the possibility of copying a mosaic, made on a large 
scale, and reproducing this as fine as desired, over a large area, by the use of a 
step-and-repeat camera. The actual printing could be done by any of the 
photomechanical processes. For the formation of a continuous roll, rotogravure 
printing could be used. 

J. H. Smith, W. Merckens, and H. B. Manissadjian (Ger. P. 197610) con- 
sidered the problem quite carefully and came to the conclusion that steel or 
copperplate engraving was the best solution of the problem. The inks were to 
be printed on paper that was coated with nitrocellulose to form a stripping 



160 HISTORY OF COLOR PHOTOGRAPHY 

paper. The inked nitrocellulose film was finally stripped on to a glass plate. 
C. L. A. Brasseur (U.S.P. 976148) came to the same conclusion, with regard to 
the printing methods. 

In printing a screen, J. S. Szczepanik (Eng. P. 6098/07) left areas completely- 
devoid of color, and had areas which were printed in black. The need for these 
is not obvious, and it would seem to be completely unnecessary, for color dilu- 
tion and unsaturation would result. The making of accurate color separations 
would also be made impossible. Other printing procedures were disclosed by 
A. Nodon (Fr. P. 372661) who printed two sets of lines horizontally, and two 
sets at an angle to these, choosing the colors so that a four-color screen with 
blue, green, orange, and red units, was formed. From the specifications it could 
not be determined whether the red and orange, and the blue and green were 
mutually exclusive in their transmissions, as would be required for a true four- 
color process. C. E. K. Mees (U.S.P. 1666048) preferred to print the screen 
elements by offset. 

The simplest type of plate with a geometric pattern is the one where the lines 
are ruled with colored inks. This is the class which the late E. J. Wall termed 
"linear screen plates," in his discussion (Wall, "History of Three-Color Pho- 
tography," p. 475). The first plates of this type were prepared by Prof. Joly, 
and independently, by J. McDonough. In 1893 Prof. Joly was issued a patent 
in which he discussed color photography in a general manner (Eng. P. 7743/93)- 
In the following year he disclosed a screen made by ruling the lines (Eng. P. 
14164/94). For the red line he used chrysoidine orange; for the green, a 
mixture of ethyl green and chrysoidine; and for the blue, water blue. The 
screen was not coated with an emulsion, but was used as the front element of a 
bipack whose rear was a panchromatic plate. Positives, made by contact from 
such negatives, were then registered with the screen. Machinery for ruling 
the colored lines was disclosed in English patents 7671/96, 8114/96, 8441/96, 
9920/96, 18097/96, and 17900/97. 

Following closely was J. McDonough (U.S.P. 561685, 561686, 561687; 
Eng. P. 12645/96). He also used the screen separate from the emulsion, and 
he suggested placing patterns in the corners of the plate to facilitate the regis- 
tration of the positives with the viewing screens. Another modification of 
dubious value was to use more than three basic colors. One of his illustrations 
showed a pattern with six designations, R, 0, F, G, J5, and I, evidently meaning 
red, orange, yellow, green, blue, and indigo. If these colors really represented 
a six-partite division of the spectrum of white light, the disclosure would have 
represented a true physical approach to the problem. But in that case, the 
description of the colors would have been in terms of spectral distribution. 
The yellow, for instance, would have represented a limited section of the band 
of colors whose limits would be 535 to 586 m/x. Orange would extend from 
586 to 647, and the red beyond this. The green would extend from 492 
to 535. Yellow filters, and in fact all yellow colors as we meet them in 
nature, transmit freely beyond 500 mju. It is feared that the image behind 



FORMATION OF THE SCREEN 161 

the yellow line in McDonough's six-line screen would be the sum of the images 
as they appeared behind the green, orange, and red lines. Machines for ruling 
the lines were disclosed by D. K. Tripp (Eng. P. 12922/01, and 15017/01), 
and by E. E. Flora (U.S.P. 676943 and 679070). It is hard to tell whether the 
English patents were issued to the original inventor, or to his English attorney, 
the English custom allowing anyone to whom a disclosure has been made to 
make application for a patent. The American custom is more rigid, so that 
only the inventor can apply for a patent. Paper prints made by the screen 
method were disclosed by McDonough in United States patent 61 1457, and in 
English patent 20417/98. 

Not very much work seems to have been done along these lines, probably 
due to the fact that it was quite difficult to rule lines in exact juxtaposition. 
In 1930 J. N. Goldsmith, a member of the Spicers organization, which was 
interested in the Dufay disclosures, and which played a prominent part in the 
exploitation of Dufaycolor, proposed to rule lines with inks. The Dufay dis- 
closures dealt mainly with the ruling of resist lines upon dyed collodion. 
Goldsmith proposed that the resist inks also contain a color (U.S.P. 1877658; 
Eng. P. 334243) which was capable of staining the collodion. The resist con- 
sisted of boiled linseed oil, or some other thickened vegetable drying oil. Red 
lines could be drawn by using an ink which contained 

Williams oil red , 20 parts 

"Thickened linseed oil 20 parts 

Acetic acid 20 parts 

Blue lines would be obtained from 

Williams oil blue 10 parts 

Thickened linseed oil 18 parts 

Acetic acid 10 parts 

A variation of this technique would be to put a dye component into the film 
base, then rule this with an ink that contained an ingredient which would unite 
with the other, to form a deep color of the desired primary. The ruled line 
would have to contain a sufficient concentration of the color-former to com- 
pletely use up the ingredient present in the base. A second ruling, this time 
with an ink containing a component which forms a different primary color, 
could be made at right angles to the first. This creates a mosaic containing 
two colors, -separated by clear areas which still contain color-forming ingredi- 
ents. Treatment of the film with a substance capable of forming the third 
primary color, completes the screen. The general idea for this is disclosed in 
United States patent 1818927, issued to J. N. Goldsmith, T. T. Baker, and 
C. Bonamico, all of the Spicers organization. Specific details are disclosed in 
English patent 352949. A film base is impregnated for several minutes in an 
alcoholic-sodium hydroxide solution containing beta naphthol. The excess is 
removed by means of a roller. If this be ruled with diazotized alpha naphth- 



162 HISTORY OF COLOR PHOTOGRAPHY 

ylamine, a series of red lines will be formed. At right angles to this can be 
ruled diazotized dianisidine to give blue lines. 

In the preferred form, the film is impregnated with Naphthol A. S. This 
is the anilide of 2 : 3 oxy-naphthoic acid. If cellulose acetate is used as the base 
material, .it is necessary to make the solution alkaline with one per cent alco- 
holic potassium hydroxide, otherwise the film will not take on the naphthol 
compound evenly. The excess is removed by means of a roller. A series of 
parallel lines is then ruled, using for the ink a strong solution of diazotized 4- 
amino-4'-ethoxy di-phenyl-amine dissolved in 50 per cent alcohol, and buffered 
with sodium acetate. After three minutes the excess is removed and a second 
series of lines is ruled, perpendicular to the first. The ink this time is a simi- 
larly constituted solution containing 2 : 5 di-chloro-benzene-diazonium chloride 
plus aluminum sulphate. The third color is obtained by dyeing the whole film 
with auramine. It is to be hoped that this is not the only ingredient of the 
third color, since that would indicate that yellow was one of the fundamental* 
units, which would be entirely wrong. 

The general idea of having a color-forming substance present in the base, 
then ruling it with a reactive material capable of color formation, is not original 
with Goldsmith and his co-workers. It had been previously suggested by 
F. Faupel in 1907 (Ger. P. 221 231). He coated a base with a colloid that 
contained ferricyanide. Upon this he ruled lines of ferrous salts, forming a 
series of lines in iron blue. This, of course, is cyan and is more fitted for a 
secondary than a primary color. But the shade of the color formed has no 
relationship to the disclosure of the technique involved. He also suggested 
that a colored untanned colloid be coated upon a film base, upon which there 
can be ruled lines with an ink containing chrome alum, tannic acid, formalde- 
hyde, or other agents which would tan the colloid. These would fix the coloring 
matter within the ruled spaces. The remaining color would be washed out, 
then replaced, and the procedure repeated after shifting the position of the 
lines. 

M. Petzold (Ger. P. 279932) ruled lines of acid dyes plus an alkaline chromate, 
upon a gelatin surface. In this manner, colored lines were formed upon the 
surface, and these areas simultaneously tanned, so that they would no longer 
absorb a pinatype dye, or a dye such as rosaniline blue. Two sets of lines, red 
and green, could be ruled at an angle to each other, the third color being put 
on by use of pinatype dyes. 

Perhaps the following should not be included in this section, but the ruling 
of lines carrying inks is one step in the procedure. It is the invention of 
A. Weckfort (Eng. P. 393797 and 402231). A series of blue lines in a greasy 
ink, is ruled on the surface of a color-blind emulsion. The ink base must be 
such that it will resist the action of water and sensitizing dye solvents. The 
clear areas are next sensitized to the green. After this, a series of green lines 
is ruled, also with a resist ink. The uncovered spaces are then sensitized to 
the red, and finally dyed red. In the meantime the coloring matter in the 



FORMATION OF THE SCREEN 163 

resist lines transfers to the gelatin surface. The removal of the resists com- 
pletes the screen plate, which now consists of a pattern of parallel lines of blue, 
interspersed with red and green squares. Directly beneath each color element 
is an emulsion that is sensitive only to the color transmitted by that element. 
M. Zeller (U.S.P. 2009424 and 2061 182), and H. Wilms (Eng. P. 436723) dis- 
closed similar procedures. These are but a modification of the O'Grady dis- 
closures (U.S.P. 1402371) who ruled the screen elements on the surface of a* 
panchromatic emulsion. After exposure, the lines were removed by washing. 
Positives made from the negative by contact were finally ruled in exact registry 
with a new set of colored lines. This presented a rather intricate registry 
problem. 

The most successful methods for the preparation of screens have been those 
utilizing resists. In these schemes, the screen base is given a partial coating 
with a resist medium. The spaces between the resist areas are then dyed. 
The first to utilize such a procedure was A. Baumgartner (Eng. P. 22138/95). 
He printed red dots with lithographic inks, and these acted as resists against 
the further action of certain solutions. The areas between the dots were next 
dyed cyan. By this means, the entire surface became coated with red (ma- 
genta) and cyan dots. On top of this was printed in a haphazard fashion, a 
yellow dot. Where the yellow overlapped the red, an orange was formed, and 
where it overlapped the cyan, a green was obtained. 

The color analysis obtained by a procedure of this type could not be accurate. 
Since the "red" became converted into an orange by the overlay of a yellow, 
it must transmit blue as well as red, so that in reality its color is magenta. The 
screen contains therefore four sets of dots, magenta, red, green, and blue-green. 
The blue primary would register under the magenta "red" and the blue-green 
dots simultaneously with the red and green primaries, respectively. The other 
primaries will also register together with one other color, under at least one of 
the four dots. No definite color separation results. Although there is no doubt 
but that a pleasing transparency can be obtained by this method, the subse- 
quent separation of this into the proper negative separations, will be practically 
impossible to make. The same error, somewhat aggravated by the fact that a 
four-color process is claimed, is made by E. Scherpel (Ger. P. 375259, 375260, 
375261). The process he used was to spray a colorless support with a resist, 
then bathe with a blue dye. The resist was removed and these portions dyed 
red. On top of this was sprayed a yellow dye which converted the red into an 
orange wherever the two overlapped, and the blue with a green where these 
and the yellow overlapped. Thus red, orange, green, and blue dots were 
formed. Of course the same objections hold here, that were noted in the 
Baumgartner case. 

G. S. Whitfield (Eng. P. 9004/06) dyed the collodion base superficially, then 
applied a resist in the form of a spray of rubber, wax, or like material. The 
areas between the resist particles were then treated with a solution that 
bleached and removed the dye. A re-dyeing with a second color was followed 



1 64 HISTORY OF COLOR PHOTOGRAPHY 

by another application of a resist. Presumedly the first spray covered but one- 
third of the total surface area, and the second spray covered one-half of the 
remaining exposed sections. The color was again removed, and the third 
primary applied. By this means it became possible to completely cover a sur- 
face with three colored areas, so that no overlaps and no undyed portions 
remain. 

Instead of using chemical agents to destroy or remove dye from a collodion 
surface, du Hauron and R. de Bercegol (Eng. P. 194/07; Fr. P. 387828; 
Ger. P. 218323 ; U.S.P. 995405 and 1005644) remove it by cutting it out. Thus 
a film base was given a coating of gelatin, and on top of this, a coating of green 
varnish. Through this last, is ruled a series of lines which cut completely 
through the varnish. Thus the dye was removed and the gelatin at these points 
laid bare. This now underwent a dyeing operation, after which a new coating 
of varnish was applied, this time colorless. The new lines were drawn through 
the varnish, cutting deeper into the gelatin so that the previously dyed por- 
tions which lay in the path of the new lines were removed. Now the third 
color was applied. Or, superficially dyed celluloid was coated with gelatin. 
This was cut away in the form of a series of lines, the cuts extending to the 
celluloid layer below, and removing the dye. A new dye, in acetone or alcohol, 
was applied. This did not stain the gelatin, since the dye solvent was not 
absorbed by it. Considerably later, after du Hauron died, R. de Bercegol 
amplified this system. A sheet of film was colored an orange red. On this was 
spread a layer of wax and a series of parallel lines cut through the wax into the 
celluloid, so that the superficial dye in the celluloid was removed. These 
portions were then colored green, rewaxed and cut again, this time with a 
series of lines at an angle to the first. The new lines were colored blue (Eng. P. 
189813; Fr. P. 554912 and addition 27364 to this; Ger. P. 392749; U.S.P. 
1673349 and 1706774). Machines for the ruling of these lines, were disclosed 
in French patent 557972, addition 27765. A similar procedure was outlined 
by I. Kitsee (Eng. P. 225659; U.S.P. 1449417, 1477880, 1477881 and 1477882). 
This time gelatin is used for the resist. The dyed celluloid is first coated with 
a layer of gelatin, and by means of a special cutting device (Fr. P. 672827; 
U.S.P. 1477883) a series of broken lines, some running at right angles to the 
others, are cut through to the celluloid below, removing the dyed portions. 
The lines are re -dyed, and the procedure repeated a second time to put in the 
third set of colors. 

The cutting method is varied somewhat by the Vereinigte Kunstseide 
fabriken (Eng. P. 21739/08 and 21840/08; Ger. P. 218298 and 230387; 
Fr. P. 395165), and independently by A. Lehner (U.S.P. 1112540 and 1112541). 
A celluloid film is grooved, and upon the raised portion is printed a greasy 
resist. This allows the dyeing of the cut portions. The resist is removed, a 
new series of grooves etched in so that the dye in the first set is removed when- 
ever the new and the old cross. A second dyeing is made. The resist is re- 
moved again, and a layer of dichromated gelatin put on. Presumedly the first 






FORMATION OF THE SCREEN 165 

two colors were red and green, so that it becomes possible to expose through 
these to form tanned gelatin corresponding to those portions of the surface 
which were still uncolored. The tanned gelatin portions are dyed up. By 
the application of heat and pressure, the screen surface was made flat. A 
similar idea is utilized by C. Spath (Ger. P. 239486; Fr. P. 399676 and 421199; 
Eng. P. 23138/10; U.S.P. 1069039 and 1108341). But this time rubber is 
used as the resist material. A series of parallel lines is ruled through the resist, 
laying bare the celluloid. This is dyed by treatment with a two per cent 
alcoholic solution of Victoria Blue. Now a new set of lines is drawn at right 
angles, and the new portions dyed with a mixture of ethyl green and auramine 
in 80 per cent alcohol. The resist is removed and the remaining portions are 
dyed with a mixture of rubin and auramine in 60 per cent alcohol. 

The Lumieres are known for the successful marketing of the Autochrome 
plate, in which the screen elements consisted of dyed starch grains applied to 
a surface by a dusting-on process. But they also did some work with geo- 
metric patterns. A resist was printed or ruled on a dyed gelatin surface (Eng. 
P. 20111/08; Fr. P. 393296; Ger. P. 207750; U.S.P. 916467). The spaces 
between the resist areas were then subjected to the action of a solution which 
destroyed the dye. This was re-dyed and the process repeated until the entire 
area was covered with a series of three-color elements. In a series of patents 
issued somewhat later, the procedure is extended (Eng. P. 29273/09, 4212/10 
and 5377/10). Two-thirds of the area of a film surface is covered with a resist. 
The intervening space is dyed with eosine scarlet. The resist is removed, and 
a new series is put on which covers one-half of the undyed areas. , The remain- 
ing half is dyed with cyanin V and metanil yellow to which have been added 
iron salts. These are present to prevent the dyes from affecting the areas 
previously dyed with eosine scarlet. The resist is removed, and this space is 
dyed with methyl violet. 

At about the same time that this series of patents was being issued, R. Krayn 
also disclosed the use of iron salts in the dye solutions (Eng. P. 26911/09; 
Fr. P. 409397; Ger. P. 221727; U.S.P. 1055189). A gelatin surface was 
printed with a greasy resist, and the intervening areas colored with cinnabar 
scarlet, then fixed with ferric chloride. The first resist was removed, and a , 
new one printed at right angles to it. The exposed surface was then dyed with 
methyl blue and fixed with ferric chloride. The resist was removed, and the 
remaining areas dyed with a mixture of patent blue A and yellow F. The 
fixing operation with ferric chloride was again applied. E. J. Wall (" History of 
Three-Color Photography," 1925, p. 494) reported that the red areas trans- 
mitted light having wavelengths greater than 560 mju, the green areas trans- 
mitted in the range from 480 to 570 m/i, while the blue transmitted only below 
490 m/x. These are ideal limits to the spectral transmissions of the three- 
color primaries in color reproduction. A "slight variation to this technique is 
noted by F. May (Eng. P. 201234). The dyes were fixed by treatment with 
ferric chloride, aluminum acetate, or formalin. 



166 HISTORY OF COLOR PHOTOGRAPHY 

A slightly different procedure was adopted by J. Rheinberg (Eng. P. 9929/14; 
Fr. P. 476228, addition 19917, and 478117; Ger. P. 326710 and 326711; U.S.P. 
1161731 and 1191034). A collodion base is dyed red, then coated with albumin 
containing *ji per cent each of ferric ammonium citrate and uranyl nitrate. 
This was exposed to a line screen having transparent and opaque lines of equal 
width. A bath in acid alcohol removed the red dye from the portions that 
had received a light exposure. These portions were then stained green. A 
resensitization with iron and uranium followed. The plate was now exposed 
to a screen at right angles to the previous lines, and this time the opaque lines 
were twice the width of the transparent ones. A wash in acid alcohol removed 
the dye in the portions that received an exposure, and these were finally dyed 
a blue. 

The Finlay plate, which enjoyed a fair measure of success up to the advent 
of the latest edition of Dufaycolor, was prepared by the use of dichromated 
gelatin. But, like the Lumieres, Mr. Finlay also did some work with resists 
(Eng. P. 421771). A metal foil is coated with dyed collodion, then a series of 
resist lines is printed upon it. Treatment with alcohol removes the green dye 
from the interspaces, and these are dyed red. A new series of resist lines is 
then printed at right angles to the first. The red dye between the resist lines 
is removed by alcohol, and the spaces re-dyed blue. After completion of the 
screen formation, the collodion surface is stripped from the metal and trans- 
ferred to a film base. 

The most successful of all the screen processes was the one initiated by 
Louis Dufay. Today the product is known as Dufaycolor, but it was first 
introduced about 1910 as the Dioptichrome plate. The first Dufay patents 
were assigned to an organization carrying the quaint name "A Company for 
the Exploitation of the Process in Color Photography of L. Dufay." This 
became the Versicolor organization. Some time later Spicers Limited became 
interested and such companies as Spicers-Dufay, Ltd., Dufaycolor, Ltd., and 
Dufay-Chromex, Ltd., were formed to exploit the disclosures. Finally the 
Ilford company became interested. In the United States there was but one 
organization, Dufaycolor, which marketed the product. 

The first Dufay disclosures (Eng. P. 11698/08; Ger. P. 237755; U.S.P. 
1003720) were concerning the preparation of the Dioptichrome plate, which 
combined dichromate printing and transfer a la wash-off-relief, for the forma- 
tion of the screen. A variation of this was contained in English patent 18744/08. 
In another disclosure, the cutting through of a resist was adopted, thus follow- 
ing du Hauron and R. de Bercegol (Fr. P. 370956, addition 7138, 388616, 
additions 10541, 13132, and 13775; 408552, additions 11696, 118454, 442881, 
addition 19753; Eng. P. 15027/12 and 27708/12; Ger. P. 273629; U.S.P. 
1 1 55900). This last group of patents was already assigned to Versicolor. 

The modern Dufaycolor film is prepared in accordance with the disclosures 
in L. Dufay's English patent 217557 and United States patent 1552126, also 
assigned to Versicolor. A celluloid film is stained superficially with a "blue" 



FORMATION OF THE SCREEN 167 

dye, and on this is printed a series of parallel lines in a greasy ink. When this 
is treated with an alcoholic solution of a "red" dye, the "blue" is removed and 
replaced by the "red." This is accomplished instantaneously if the second 
solution is made alkaline. After the resist is removed, the entire film is dyed 
yellow. This converts the "blue" and "red" into green and orange, this in- 
dicating that the original colors were really cyan and magenta. A new series 
of resist lines is now printed at right angles to the first, and the places free 
from resist are treated with a blue-violet dye solution. This will remove the 
yellow dye completely, and the cyan and magenta but partially. These will 
become converted into a blue-violet or red-violet respectively. Dufay claims 
a four-color process here, but it appears a rather dubious claim, since the last 
two colors would have a tremendous overlap, hence are not really independent 
of each other. As an example, the first dyeing is made with methylene blue. 
The second is made with Rhodamine B. The yellow overall is made with aura- 
mine. At this stage the lines will be true primary red and green. After the 
final operation, the yellow becomes replaced with crystal violet, a dye that 
transmits much more blue than red, but still a significant amount of that pri- 
mary. This on top of Rhodamine B will give a reddish violet, a color that 
will transmit red and blue freely, but no green. This certainly cannot be con- 
sidered as a unit in a four-color process, since the same colors will be registered 
simultaneously under the red and the blue elements. These last will be formed 
at those places where crystal violet overlays the methylene blue. The first will 
prevent any green rays from being transmitted, while the second will prevent 
the red rays from being transmitted. But both will transmit the pure blue. 
It is interesting to note that every claim made so far by the inventors of four- 
and more color processes, fails when analyzed in this manner. They only suc- 
ceeded in destroying the purity of the color analysis. 

The last patent issued to L. Dufay was the United States patent 1805361, 
issued in 1 93 1, and assigned to H. Wade. The corresponding English patent 
322432 was issued to H. Wade and assigned to Versicolor, a rather curious 
state of affairs. The American patent introduces a few names which later 
play a prominent part in the further development of Dufaycolor. The film 
base is prepared in accordance to the disclosures of H. J. Hands (Eng. P. 
279139, 281803 an d 294008). This is coated with a layer of collodion, dyed 
green by the addition of 2 cc of a stock dye solution to 30 cc of the collodion. 
This is a development of C. Bonami^o (Eng. P. 321222). The stock dye solu- 
tion has the following composition: 

Malachite green 4 parts 

Auramine 6.7 parts 

Alcohol 100 parts 

About 25 cc of collodion is used to coat an area measuring 100 by 25 centi- 
meters, yielding a layer that is 0.01 mm thick when wet, and from 0.0002 to 
0.0005 mm when dry. 



168 HISTORY OF COLOR PHOTOGRAPHY 

On this is ruled a series of lines with a resist ink. This was done in accord- 
ance with the disclosures of H. Wade (Eng. P. 322454; U.S.P. 1760048). 
Fifteen lines were drawn to the linear millimeter by means of a steel roller, at 
an angle of 23 degrees to the axis of the cylinder. This corresponds to 400 
lines to the inch. The present material is somewhat finer, containing approxi- 
mately 500 lines to the inch, and some screens were made in this manner with 
750 lines to the linear inch. The ruled lines were dried for one hour, and then 
bathed in a solution whose composition was as follows: 

Alcohol 100 parts 

Aqueous KOH, 10% 2 parts 

Acetone 4 parts 

Treatment in this solution removed the dye from the clear spaces, which are 
re-dyed by passing over a dye roller. The film base was thoroughly washed to 
remove excess dye, then it was passed through a solvent which removed the 
resist lines. A buffing action prepared the surface for a new set of lines, ruled 
at right angles to the old. The spaces clear of ink lines were again decolorized 
by the action of the alkaline-acetone-alcohol solution, and the un-dyed por- 
tions dyed blue by means of the following bath: 

Crystal violet, 4% alcoholic solution 80 parts 

Malachite green, 8% alcoholic solution 20 parts 

The film base was cleared as before, then coated with a panchromatic emulsion. 
The further development of the Dufaycolor material was carried on by J. N. 
Goldsmith, T. T. Baker, and C. Bonamico. 

Although Dufay no longer contributed to the solution, the procedure was 
not changed. Rather, all efforts were centered upon the improvement of de- 
tails. The first serious problem was that of putting a superficial coating of 
dye upon the film base, before the resists were applied. This coating must be 
absolutely uniform, and of sufficient density to act as a complete filter for one 
primary. The thickness of the coating must be kept to a minimum, otherwise 
the slight divergence of the rays after passing through the screen will cause 
the deposition of densities of one color under the areas intended for the other 
two. It is the screen which is in the rear focal plane of the lens system, so 
that divergence of the rays starts in the plane of the screen. Acid dyes do not 
stain collodion, so that the screen colors are limited to the basic dyes. But 
celluloid, which is collodion to which camphor and other plasticizers have been 
added, will not stain when treated with basic dyes. It seems that the addition 
of the plasticizer extensively changes the properties of the collodion. 

A whole series of patents, all issued jointly to the trio J. N. Goldsmith, T. T. 
Baker, and C. Bonamico, deal with the treatment of celluloid film to make it 
receptive of basic dyes. It is to be recalled that the basic patent (Eng. P. 
322432) utilized a film base of cellulose acetate, and that this was coated with 
a layer of collodion dyed green. The first disclosure in which the attempt is 



FORMATION OF TEE SCREEN 169 

made to prepare the film base itself for a superficial dyeing, is in English patent 
333865. The base is treated with an alcoholic potash solution, ranging from 
one-half to five per cent of the lye. After this treatment, it will take on dye 
uniformly from a bath such as 

Malachite green 2 parts 

Alcohol 60 parts 

Water 40 parts 

Acetic acid 4 parts 

If the film base is celluloid, then this is coated with a thin layer of cellulose 
acetate, which is deposited from the following solution: 



Cellulose acetate 


1 part 


Tetra-chlor-ethane 


20 parts 


Methyl alcohol 


2§ parts 


Acetone 


z\ parts 



A coating such as this readily takes up basic dyes from solutions that would 
have no effect whatsoever upon celluloid. Therefore the coated film base 
could receive its superficial dyeing on one side with the absolute assurance 
that the other side would remain clear. This represents a tremendous simpli- 
fication in the manipulations of the film base during the period of the screen 
formation. In a later patent there is disclosed still another scheme for putting 
a superficial coating of cellulose acetate on the film base (Eng. P. 337073). 
The film passed around a roller which just dips into a solution of cellulose ace- 
tate in acetone, such as 

Cellulose acetate 14-15 parts 

Plasticizer 5-6 parts 

Acetone % 100 parts 

The above treatment yields a layer which has a thickness of 4 to 8 ju (0.004 to 
0.008 mm). Such a coating will not absorb the dye, but it can be given a sur- 
face coating of collodion which is already dyed. The cellulose acetate evidently 
acts as a subbing base for the dyed collodion layer. 

The various difficulties encountered in the making of a uniformly dyed sur- 
face layer are discussed in English patent 339238. It is very difficult to apply 
a dye to a celluloid film, but once applied, the dye is retained with great tenac- 
ity. This makes its subsequent bleaching extremely hard. This can be over- 
come by the application of a coating of pure cellulose acetate to the surface, as 
is disclosed in English patent 334265. When a film base of cellulose acetate is 
to be used, other difficulties arise. Too deep a stain is taken up, and one 
which does not discharge easily. Coating such a film layer with collodion 
(Eng.* P. 322432) helped alleviate this situation. In either case, however, the 
precautionary measures were nullified if the dye diffused beyond the super- 
ficial layer into the film base itself. To prevent this, the film may be dyed 



170 HISTORY OF COLOR PHOTOGRAPHY 

from a solution which is non-penetrating. This is accomplished by dispersing 
or dissolving the dye in castor oil, castor-oil-alcohol, or gum arabic mixtures. 

To insure that the sound-track areas of motion picture film would be left 
free of screen pattern, C. Bonamico (U.S.P. 2008239; Eng. P. 356816 and 
414761), and T. T. Baker (Eng. P. 335899) so constructed the rollers which 
printed the resists that no deposits were made in those areas which, after 
slitting, would correspond to the sound-track region. 

The next problem to be tackled was that of coating. Here it was necessary 
to insure the firm adhesion of a gelatin layer to another of rather complex struc- 
ture, and one which contained basic dyes. These were not always neutral 
with regard to their action upon the photographic - emulsion. Some of the 
dyes may act as sensitizers, while others may have the opposite action and act 
as desensitizers. So the simplest procedure to adopt was to insulate the 
screen layer from the emulsion. This was disclosed by T. T. Baker (Eng. P. 
401719; U.S.P. 1962679) and by H. D. Murray, H. Baines, and R. A. S. 
Grist (Eng. P. 435484). A solution of resin in benzol to which some linseed oil 
had been added, was the Baker solution. 

A complete specification for the formation of the screen is disclosed by 
T. T. Baker (Eng. P. 420824; U.S.P. 2030163). The reason for this patent is 
to protect the printing of the resist lines by means of a special roller, where non- 
printing areas contain an ink-resisting mercury-silver amalgam. The film base 
is prepared in the manner disclosed by H. J. Hands (Eng. P. 243032, 281803, 
287635, and 301439). On this is coated a thin layer of collodion which is dyed 
green by the addition of 2 cc of a stock dye to 30 cc of collodion. The dye is 
prepared by dissolving 4 parts of malachite green and 6.7 parts of auramine 
in 100 parts of methyl alcohol. A strip 100 by 26 centimeters requires from 
22 to 25 cc of the dyed collodion solution. This had been disclosed previously 
by C. Bonamico and H. Wade (Eng. P. 321222). Upon the dyed film is then 
printed a series of resist areas, which may or may not be in the form of a geo- 
metric pattern, by means of a special roller. 

This is prepared in the following manner. A metal roller is first given a 
coating of silver. Upon this is deposited an even coating of chromium, also 
electrolytically. On top of the chromium is placed a sensitized tissue upon 
which the screen pattern has been photographically printed. The tissue used 
is a dichromated gelatin or other colloid. It is placed with the image portion 
adjacent to the surface of the roller. Treatment with hot water removes the 
non-image portions, leaving a tanned colloid relief image upon the surface of 
the roller. This is next subjected to the action of a solution which will etch 
the chromium, wherever this surface has been laid bare, obviously between the 
tanned colloid relief areas. When the chromium surface has been etched 
away, leaving bare the silver, the etch is stopped. The tanned colloid resist is 
removed, and the roller treated with a mercury solution. Only the silver will 
form an amalgam, the chromium being insoluble in mercury. An amalgam 
surface has the extremely useful property of acting as a resist to greasy inks, 



FORMATION OF THE SCREEN 171 

while the chromium surface will adsorb the ink. Thus a matrix is formed by 
means of which a resist containing a superficial coating of dyed collodion can 
be transferred to the cellulose acetate film. This technique is a well established 
one in the photomechanical printing industry, especially in that section which 
uses planographic printing with mercury inks. 

After the resist has been transferred, the dye in the areas between the lines 
is bleached by treatment with 

Methyl alcohol 100 parts 

Caustic potash, 2 % solution 2 parts 
Acetone 4 parts 

The bleached areas are then re-dyed by treatment with an 8 per cent alcoholic 
solution of basic red N. The old resist lines are removed by treatment with 
a suitable solvent, and a new set of lines printed at right angles to the first. 
Decoloration is again accomplished in the areas between the greasy lines, this 
time both red and green being removed. At this stage the film base consists of 
a series of squares of red and green lying in juxtaposition, and running in lines. 
The lines are separated from each other by clear undyed lines that are free 
from grease, and which can therefore be stained from alcoholic solutions. This 
is accomplished by means of a blue bath, whose composition is the same as the 
one given above (cf. p. 168). A wash removes excess dye, while treatment with 
a solvent removes the greasy lines. The screen elements are then coated with 
an insulating layer of varnish, which also serves as a subbing for the pan- 
chromatic emulsion which is coated upon the varnish layer. 

There are several schemes which cannot readily be classified in any group. 
F. J. H. Harrison (U.S.P. 578147) used a band of celluloid that was colored 
solidly, and ruled with opaque lines which were twice the width of the trans- 
parent spaces between. The celluloid "ribbon" was divided into three sec- 
tions, each of which was stained in a different primary color. The lines in 
each section were staggered with respect to those in adjacent sections. The 
ribbon was moved by means of a spring motor across the front of the sensitive 
plate. The result was the same as if the plate were exposed behind a screen 
composed of colored lines lying in juxtaposition. It is hard to determine just 
exactly what advantage there was in this procedure outside of the fact that it 
is quite simple to prepare such a screened ribbon, for the subsequent separation 
or the preparation of a colored transparency would impose a very ticklish 
problem of registry. Similar ideas were expounded by C. L. A. Brasseur and 
S. P. Sampolo (Eng. P. 8390/96) ; and by F. E. Ives (U.S.P. 648748 and 666424). 
In the Ives procedure, the black-and-white screen was placed in front of the 
plate, and the three exposures were made through the three filters. Between 
the exposures, the position of the screen was shifted, so that a new area was 
exposed on the film each time, and through a different filter. 

A rather simple procedure was adopted by M. Obergassner (Ger. P. 263819; 
Fr. P. 438746). A plate was coated with dyed gelatin. It was then grooved 



172 HISTORY OF COLOR PHOTOGRAPHY 

through to the glass in a series of parallel lines. After recoating with another 
layer of dyed gelatin, a new set of grooves was cut through. The plate was 
recoated a second time, buffed even, then coated with a panchromatic emulsion. 
It is to be recalled that the earliest patents of du Hauron and Bercegol utilized 
a technique identical to this, except that they cut grooves through a resist 
surface. This, therefore, is a simplification of the earlier disclosure. C. L. A. 
Brasseur (Eng. P. 20909/08; Ger. P. 219977; U.S.P. 976118) combined photo- 
mechanical printing methods with the dusting-on procedure. He printed a 
tacky material upon a glass or film base, then dusted on colored powders. 
This was followed by another printing and dusting, until the film surface was 
completely covered. F. Faulstich also combined two methods (Eng. P. 
152002). A base was sprayed with a dye so that only part of the total surface 
was covered, liie covered areas then acted as resists for a further dyeing 
operation. Somewhat similar in basic principle was the technique adopted by 
Keller (Eng. P. 244644). He sprayed a gelatin-coated base with colors com- 
pounded to contain alum or formaldehyde. These chemicals tanned the 
gelatin in situ with the dye powder, making that specific area no longer recep- 
tive of dyes, if these were of the pinatype group. 



CHAPTER 14 
PROCESSING SCREEN PLATES 

(A). Hypersensitization. — The early screen plates were extremely slow, 
from thirty to sixty times as slow as normal emulsions of their period. This 
lack of speed made itself felt very soon after their introduction, and much 
effort was expended in attempts to overcome it. One type of effort was to 
hypersensitize the emulsion just before use. At first the attempts were con- 
fined to an increase of the red sensitivity, but soon the attention became 
centered upon a general intensification of the plate sensitivity. No doubt the 
latter efforts were directly the results of the previous ones. Optical sensitiza- 
tion proceeds best when carried out in an alkaline medium, and this condition 
also leads to a large increase of general sensitivity. Later, when the screen- 
type materials could be made only a little slower than the normal, it was no 
longer necessary to resort to hypersensitization, but the discussion of this 
subject is included here mainly because of historical interest. 

The photographic emulsion consists of a relatively coarse dispersion of micro 
crystals of silver iodo-bromide in gelatin. The pure emulsion is sensitive only 
to the blue and violet portions of the spectrum. In 1873, H. Vogel dis- 
covered that when the silver halide grain was treated with certain dyes, the 
sensitivity of the grain became extended to the spectral regions absorbed by the 
dye. His discovery, accepted several years after the first disclosure, soon un- 
loosed a veritable barrage of work, and after many years it became possible to 
prepare photographic emulsions whose sensitivities cover any desired range 
from the far ultraviolet, well below 200 mju, to the extreme infrared above 
1 200 mji. The visible range extends from 400 to 700 m/z. 

The first sensitizing experiments were carried out by bathing the finished 
plate in a dye solution. Plates sensitized by bathing in this manner were quite 
effective in the new range, but unfortunately had very poor keeping qualities. 
It was soon established that excellent keeping qualities would be obtained if 
the sensitizing dye were added to the emulsion just before it was coated. But 
the degree of sensitivity induced was not quite so high in this case as when the 
coated plate was bathed and dried just before use. Bathing appeared to give 
a hypersensitizing effect. This was traced to the removal of the slight amount 
of soluble halide left in the finished emulsion to give it keeping qualities and 
freedom from fog. Carroll, Hubbard, and Kretchman made a special study of 
the effect that halides had upon the general sensitivity and the keeping qual- 
ities of emulsions. Their conclusion was (Bu. Stand. J. of Res., Vol. 12 (1934), 

173 



174 HISTORY OF COLOR PHOTOGRAPHY 

p. 223) that a finished emulsion should contain approximately one to five moles 
of soluble bromide per 1000 moles of silver bromide, in order for that emulsion 
to keep over a year. The presence of even this small amount of excess bromide 
was enough to materially decrease the general sensitivity of the plate, and to 
reduce to an even greater extent, the capacity of that emulsion to be optically 
sensitized. 

When a finished plate containing the absolute minimum of soluble bromide 
requisite for good keeping qualities was bathed in a solution capable of re- 
moving some of the bromide, it was soon readily seen that the new plate was 
not only faster and more sensitive in the induced range, but that it also had 
poor keeping qualities. Efforts to replace the bromide resulted merely in a 
reduction of color sensitivity and general speed, although the keeping qualities 
did improve. This lack of stability was the stumbling block that lay in the 
path of the successful utilization of the Friedman Bipacks (cf. chapter on 
Monopacks). 

A considerable intensification of the plate speed can be obtained by making 
the bath alkaline with ammonia. For a long time, bathing a plate in an 
ammoniacal solution was the recognized technique for hypersensitization. This 
action was first disclosed by Schumann (Phot. Woch. (1885), p. 395; (1886) 
p. 49). It was soon verified by others, including Walters and Davis (Bu. 
Stand. Sci. Paper No. 422), Burka (/. Frank. Inst., Vol. 189 (1920), p. 25) and 
Jacobsohn ("Theorie und Praxis der Ubersensibilisierung"), who wrote a 
complete review that covered most of the work done prior to the contributions 
of Carroll and Hubbard (Bu. Stand. J. of Res., Vol. 10 (1933), p. 211). 

These men, applying the Donnan equilibrium relationship, have shown that 
the net effect of bathing a photographic emulsion with ammonia was to create 
an excess of silver ions in the emulsion, a condition that is very prone to fog as 
well as to increased speed. This they proved experimentally, using Seed 23 
plates. The amount of excess silver ions ranged from 0.2 to 1.0 per cent. The 
same effect could be obtained if the plate were treated with silver solutions, 
hence the hypersensitizing action of solutions containing silver ions could be 
explained by exactly the same mechanism. It is a well-known fact that the 
basic sensitizers act much better in the absence of soluble halides, than in their 
presence. Hence it is to be expected that a plate sensitized with basic dyes 
will show an increased activity by treatment with soluble silver salts, which 
will remove the excess bromide ions completely, and leave the plate with excess 
silver ions. Bokinik (Zeit. Wiss. Phot., Vol. 30 (1932), p. 322) reported that if 
a plate were dyed with a basic sensitizer, dried, then bathed with dilute silver 
nitrate solution, a considerable increase in the effective sensitization was ob- 
tained. Even some of the desensitizers gave definite sensitizing action under 
these conditions. 

As soon as it was fully realized that the presence of halide ions affected the 
sensitivity of the plate to a great extent, efforts were directed to reduce this 
concentration to an absolute minimum, not by treatment with silver nitrate, 



PROCESSING SCREEN PLATES 175 

which would leave excess silver ions, but by treatment with silver chloride 
solutions. One of the first to do this was F. Monpillard (Bull. Soc.franq. Phot., 
Vol. 64 (1922), pp. 90, 130). The addition of silver ions to fluorescein sensitizing 
solutions has been known at least since 1884, when W. Abney suggested that 
the silver salt of eosine be added to an emulsion to give it orthochromatic 
properties (Phot. News, Vol. 28 (1884), p. 500). A short time later W. H. 
Hyslop substituted an ammoniacal solution of pure silver chloride for the silver 
nitrate (Phot. News, Vol. 31 (1887), p. 107) in the preparation of the silver salt 
of the sensitizer. Silver nitrate (3.6 grams) was treated with excess hydro- 
chloric acid. The precipitated silver chloride was washed thoroughly, then 
dissolved in concentrated ammonia and diluted with water to give a final 
volume of 10 cc. The erythrosine (5.4 grams) was meanwhile dissolved in 
3.6 cc concentrated ammonia, and then diluted to 175 cc with alcohol. From 
these stock solutions, the sensitizing bath was compounded as follows: 



Silver solution 


1.85 parts 


Dye solution 


6.3 parts 


Water to 


1000 parts 



The Monpillard technique merely substituted the cyanine dyes for erythro- 
sine. Pinaverdol (1 part), pinachrome (0.5 part) and pinacyanol (0.5 part) 
were each dissolved in 1000 parts of alcohol. From these a stock dye solution 
was made containing 160 parts each of pinaverdol and pinachrome solutions, 
80 parts of pinacyanol, and 600 parts of alcohol. A stock silver chloride solu- 
tion was also prepared by dissolving 0.2 part of the silver halide in 8 parts of 
concentrated ammonia, then making up to 100 parts with water. The sensi- 
tizing solution was made by adding 100 parts of the stock silver to 100 parts of 
the stock dye, and diluting to 1000 parts with 50 per cent alcohol. The tem- 
perature of the solution should be kept very low. After a five-minute treat- 
ment, the plates should be washed and dried as rapidly as possible. Since the 
emulsion would not keep for more than thirty-six hours, the technique was not 
commercially feasible. A thirty-fold increase in speed was claimed. 

Approximately the same idea was disclosed by A. Ninck (Bull. Soc. franq. 
Phot., Vol. 65 (1924), p. 345). He used pantochrome for the sensitizing 
agent (1:2500), and 2 per cent silver chloride containing 6.5 parts of concen- 
trated ammonia per 100 parts of solution. The actual sensitizer contained: 

Dye solution 20 parts 

Silver solution 6.6 parts 

Water to 1000 parts 

Ninck claimed a speed increase of thirty times, and a keeping quality vastly 
superior to that of Monpillard. He also studied the effect of silver concentra- 
tion (Bull. Soc. franq. Phot., Vol. 66 (1924), pp. 8^ and 92). All the tests were 
made on the Autochrome plate, the exposures being made behind a Eder- 
Hecht wedge. 



176 HISTORY OF COLOR PHOTOGRAPHY 

Test No. Silver Content in Sensitizing Bath Relative Speed 



I. 


Control 


1 


2. 


0.00 gram per liter 


6 


3- 


0.07 gram per liter 


17 


4- 


0.18 gram per liter 


23 


5- 


0.20 gram per liter 


33 


6. 


°-53 gram per liter 


40 


7- 


i.6q grams per liter 


17 



The control plate was an Autochrome which received no treatment. Test No. 2 
was made on an Autochrome which was treated with the sensitizing bath, but 
compounded to contain no silver. This test indicated the effect of bathing the 
plate in an ammoniacal dye solution. It is interesting to note that a maximum 
exists in the amount of silver that may be used. As the silver content of the 
bath increased, the keeping quality of the plate decreased. Old plates, that 
were no longer fit for use because of their tendency to fog, could be revived by 
treatment with : 



Chromic acid 


5 parts 


Potassium bromide 


10 parts 


Water to 


1000 parts 



After a five-minute wash, the plate could be hypersensitized. This treatment 
obviously destroyed the silver specks that were the cause of the latent fog. 

Many years later L. J. Meker (Bull. Soc.franq. Phot., for 1936; Phot. Ind., 
Vol. 35 (1937) p. 12) disclosed an identical procedure for the renovation of old 
fogged Autochromes. His sensitizing bath was compounded as follows: 



Pinaverdol stock 


30 parts 


Pinachrome stock 


30 parts 


Pinacyanol stock 


15 parts 


Silver stock 


1 part 


Concentrated ammonia 


3 drops 



The dye stock solutions contained one part of sensitizer to 2000 parts of 
alcohol. The silver solution was made by adding 15 parts of concentrated 
ammonia to 5 parts of a 10 per cent silver nitrate solution, and diluting with 
25 cc of water. 

Other systems of hypersensitization were based upon a reaction with the 
silver halide grain itself. It has always been felt that the latent image was 
composed of a grain of silver halide upon which there was adsorbed a speck of 
metallic silver. The brilliant photographic chemist Carey Lea had demon- 
strated that silver halide grains upon which there had been adsorbed colloidally 
dispersed silver, were developable without the action of light. But before 
developability could be conferred, a certain minimum size had to be reached 
by the adsorbed particle. It is possible to form the metallic silver particles by 
other means than by the action of light. Mild reducing agents could be 



PROCESSING SCREEN PLATES 177 

allowed to act on the grain for a period just sufficient to reduce a few molecules. 
Working along lines like these, Luppo-Cramer found that hydrogen peroxide 
intensified a latent image if the plate, after exposure, were treated with a weak 
alkaline solution of the peroxide {Phot, Korr., Vol. 52 (1915), p. 136). Sheppard 
and Wightman {J. Frank, Inst., Vol. 195 (1923), p. 337; Vol. 200 (1935), p. 
3355 Vol. 203 (1927), p. 261; Vol. 204 (1927), p. 731; Brit. J. Phot., Vol. 74 
(1927), p. 447) studied this action critically, and found that the action Would 
be obtained whether the plate was treated with peroxide before or after ex- 
posure. They found that peroxide took the place of light in creating a latent 
image. If the concentration of the peroxide was kept to a value that would 
produce an effect just sufficient to overcome the inertia of the emulsion (an 
effect equivalent to the toe of the H & D curve), then hypersensitization occurs. 
Schmieschek attempted to put this idea into practice by compounding an 
ammoniacal silver solution with peroxide present {Phot. Ind., Vol. 28 (1930) 
pp. 445, 472), but the best opinion has it that his conclusions are unfounded. 

More recently F. Dersch and H. Duerr {J. Soc. Mot. Pict. Eng.,Vo\. 28 (1937), 
p. 178) disclosed still another hypersensitizing procedure, one which had prac- 
tically no effect upon any other photographic characteristic of the emulsion 
besides decreasing the inertia. Their technique could be applied with equal 
results either before or after the exposure had taken place. It was merely 
necessary to put a drop of mercury into the container holding the film, taking 
care that no part of the emulsion came in contact with the mercury. Where 
this happens, intense fog results. Possible theories as to the mechanism of the 
reaction are contained in a discussion of this phenomenon by J. S. Friedman 
{Am. Phot., Vol. 32 (1937), p. 738; Vol. 34 (1939), p. 700). The hypersensitiza- 
tion that is obtained by the action of mercury is mild. The speed is increased, 
at most, approximately twice. This system has but little appeal, therefore, 
to screen-plate users. But it does hold out the hope for better results. 

The latent image consists of a speck of colloidally dispersed silver adsorbed 
upon a silver halide grain. But a grain that has colloidal gold, copper, or any 
one of many other metals adsorbed upon it, is also in a developable condition. 
From this point of view it can be considered that when a film is exposed to 
mercury vapor, some mercury atoms become adsorbed by the silver halide 
grains and in that way make them developable. The total number of grains 
thus affected may be just sufficient to overcome the inertia. Or the grains can 
adsorb an amount of mercury that is just insufficient to cause developability. 
It would require much less light, therefore, to bring these grains into a develop- 
able condition. In either case a reduction of the emulsion inertia takes place. 

The vapor tension of mercury at room temperature is very low, 0.00109. 
At 50 C the value is 0.0127, or .twelve times as much. If the temperature at 
which the film is exposed to the mercury is maintained at 18 C, approximately 
20 hours is required to effect a two-fold increase in speed. But at 50, only a 
few minutes treatment is required. This indicates a tremendous effect due to 
an increase in the concentration of the mercury vapor. Since the action can 



178 HISTORY OF COLOR PHOTOGRAPHY 

be carried out after exposure, it may be possible to duplicate the result by 
compounding colloidal solutions of mercury in water or other liquids, and 
subjecting the exposed film to this. 

So much for the subject of hypersensitization. As pointed out at the first 
part of the chapter, Dufaycolor, with a speed rating of Weston 8, was suffi- 
ciently fast not to require this treatment. 

(B). Chemical Reversal. — Since the registry of the image with the 
screen elements must be so exact, the preferred method of processing screen 
plates has been by reversal. This popularized a form of technique mainly 
devoted to amateur motion-picture film. Briefly described, the procedure is 
as follows: The plate, after exposure, is developed in the normal manner, but 
it is not fixed out. Instead, it is treated with acid dichromate or permanganate, 
which dissolves out the silver image. The remaining silver halide grains are 
now present in the form of a positive, since a negative image has been removed. 
Exposure to light followed by a second development, yields a positive which is 
registered exactly with the screen elements. There are several features in 
reversal development which make the procedure very useful not only for the 
processing of materials like Kodachrome, Ansco Color, Autochrome, Dufay- 
color, etc., but for all types of work. 

As has been stated so often in these pages, the photographic emulsion con- 
sists of a complex mixture of silver halide grains whose sizes and sensitivities 
vary over a wide range. In any one emulsion, the larger grains are the most 
sensitive ones. During the initial development, a group of grains come together 
in some mysterious fashion to form a single silver grain which makes up the 
final image. The image that is developed during normal procedure will there- 
fore contain grains that are fairly coarse and that are not too evenly distributed 
throughout the emulsion. But the grains that are left behind, after the initial 
image has been developed, consist of the finer grains of the emulsion; and 
these, during development, will not have as great a tendency to clump, since the 
individual particles are insulated from each other much more thoroughly. A 
reversed image is therefore universally much finer-grained and more evenly 
distributed than a direct image. 

Another advantage is that a given emulsion will have a greater speed if 
processed by reversal than if processed normally. This is usually caused by the 
fact that the first development in reversible procedure is practically always 
carried out to gamma infinity. Every single grain that has received an ex- 
posure sufficient for the formation of a latent image, is reduced. If this were 
done with normally processed plates, the developer would have to be com- 
pounded to contain a relatively high concentration of soluble bromide, other- 
wise intense fog would occur, which would ruin the printing qualities of the 
negative. But in reversal procedure, the formation of a high fog level is of no 
importance whatsoever, since the silver that is formed initially is removed. 
This holds for the fog as well as for the image itself. 

There are several disadvantages, however, in the use of this procedure. 



PROCESSING SCREEN PLATES 179 

There is first of all, a reversal of the image, as regards left and right. This 
means that the image must be copied with the base facing the emulsion side 
of the copying material. Secondly, the image is a positive transparency, and so 
before copies can be made it must be converted into a negative. In color 
photography, however, especially in the photomechanical field, it is sometimes 
necessary to use positives from which to print. In those cases, the printing 
image is one step closer to the original than would be true for the normal 
procedure. 

Reversal processing has received a considerable amount of study, and we 
cannot in these pages review all the work. But since we are concerned with the 
technique from a practical point of view, we will discuss the disclosures of P. K. 
Turner, who made a special study of the working conditions best suited for 
reversal {Brit. J. Phot., Vol. 84 (1937), pp. 435, 449, 465). The problem that 
confronts the technician is best explained by Mr. Lloyd E. Varden (J. S. Fried- 
man, Am. Phot., Vol. 33 (1938), p. 283), who writes, in substance, as follows: 
The fact that an emulsion cannot be completely exposed and developed in one 
operation, would lead to degraded highlights in reversal processes if precautions 
in the first development were not taken. An emulsion can be exposed to a 
certain limit, at which limit some of the grains begin to solarize. The first 
developer is always a strong and highly caustic type of developer which reduces 
all the grains that have been exposed, but not exposed to the point of solariza- 
tion. This development is prolonged beyond normal to make sure of this. If 
the exposure has been excessive and some of the grains became solarized, these 
grains will not be developable. Clear highlights in the final image depend upon 
the complete absence of developable silver-halide grains in these regions after 
the initial development is over. To insure this, silver-halide solvents are 
added to the solution, such as ammonia, hypo, etc., but the preferred agent is 
potassium thiocyanate. 

If a piece of film is exposed to the limit (not to solarization), and developed to 
the limit, there will still be some silver-halide grains left, which, after the re- 
moval of the developed silver, will give appreciable density. In reversal work, 
this silver-halide residue must be removed, and as stated before, this is accom- 
plished by the addition of suitable solvents to the developer. Thus simultane- 
ously with the reduction of the silver halide to metallic silver, it is also dis- 
solved away preferentially in the regions of high density. The amount of 
solvent added should be sufficient to insure the complete removal of the silver 
halide left in the highlight regions of the film. 

In his paper, Mr. Turner discusses the problem generally ; There are four 
techniques available in reversal processing. 

1. Controlled time of first development. 

2. Controlled exposure to light after reversal. 

3. Controlled second development. 

4. Use of a special type of first developer, 



180 HISTORY OF COLOR PHOTOGRAPHY 

The first method can be discarded at once, as it is only for a single-exposure 
intensity that a sufficient number of grains become affected in highlight regions 
to enable a suitable density range to be covered in the final print. This makes 
its use by far too critical to be of practical value. The second method requires 
accurate measuring devices and processing conditions that are not readily avail- 
able to the average technician. It is the method used by the Eastman Kodak 
Company. After the bleach treatment which removes the silver negative 
image, each frame in the film is metered by red light, photo-electrically, and 
then exposed automatically in accordance with this reading. It is seen that 
there must be had an accurate knowledge of the effect that the removal of the 
first image has upon the speed of the residual grains. This is complicated by 
the fact that not every frame will have the same proportion of the fast grains 
affected, so that a considerable variation of residual film speed exists. The same 
objections may be noted for case three. It is only the last technique that re- 
mains available for the average technician. This is the technique used by 
Autochrome and Dufaycolor. 

Mr. Turner showed that with a sufficient quantity of bromide present in the 
developer to reduce the speed of the emulsion to one-fourth of its original, and 
with enough hypo present to allow the fixation to proceed at the rate that is 
one-tenth that of the development, it would be possible to obtain a brightness 
range in the positive of from one to thirty. To determine these constants more 
accurately, it is necessary to know the film speed of the emulsion for normal 
development. A strip of the emulsion is exposed assuming twice the speed. 
The developer is then compounded with no bromide present, so that complete 
development is obtained in five minutes. Add 10 grams of potassium bromide 
and 1 6 grams of hypo, or 1.6 grams of potassium cyanide, per liter of solution. 
Two strips of film, with identical exposure (half normal), are developed in this 
solution, one until the highlights appear as dense when viewed through the base 
as when viewed through the emulsion; the other strip fifty per cent longer. Of 
course the two film strips are desensitized to allow visual inspection during de- 
velopment. The strips are placed in the acid dichromate reversal baths 
directly from the developer bath, then cleared in bisulphite which removes the 
dichromate stain. Redevelopment in the first developer completes the opera- 
tion. The two strips should have the same contrast. If they have different 
contrasts, the original exposure was wrong. If the strips appear too contrasty, 
more hypo should be added to the solution. If they are too soft, less hypo 
should be used. The developer suggested by Mr. Turner, is the following one: 

^Metol 25 parts 

Hydroquinone 50 parts 

Sodium sulphite 200 parts 

Sodium hydroxide 25 parts 

Water to 1000 parts 

For use, take one part of this, add one part each of 12^ per cent hypo and 10 
per cent bromide, and 7 parts of water. 



PROCESSING SCREEN PLATES 181 

The effect of adding silver halide solvents to a developer, was studied by 
H. D. Murray and D. A. Spencer {Phot. /., Vol. 77 (1937), pp. 330, 458). They 
found that in general the addition of substances like hypo or thiocyanate to a 
developer, caused an acceleration in the rate of development, if the concentra- 
tions were not too high. Mannes and Godowsky have found that the addition 
of approximately one gram of thiocyanate per liter of developer, gave a tre- 
mendous increase in the potential (U.S.P. 2091 7 13). Other silver solvents that 
can be used are ammonia and its derivatives, such as the ethanolamines, 
ethylene diamine, etc. These have certain other advantages, such as the fact 
that they can also be used instead of the carbonate or other energizers. 

The next step in the procedure is the removal of the silver image. This is 
usually called reversal, for after this step is accomplished, a positive residue 
remains, consisting of silver bromide. The removal of the silver can be ac- 
complished by treatment with acidified dichromate or permanganate. The 
acid ion must be one that forms a soluble silver salt, such as sulphate, nitrate, 
or acetate, otherwise the silver will merely be changed into an insoluble silver 
salt. This, by the action of the second developer, will become reconverted into 
silver. Sulphuric acid is generally the most desired acid ion. 

When gelatin comes in contact with silver ions, silver gelatinate is formed, a 
substance that will be easily reduced to metallic silver by the action of even very 
mild reducing agents. It is not possible to wash this substance out, so that 
regardless of how long the wash after the reversal is prolonged, the silver- 
gelatin complex will remain, and perhaps even grow. This last may be due to 
the formation of excess silver ions by the hydrolytic action of water upon silver 

bromide. 

H 2 
AgBr <=± Ag+ + Br~ 

The bromide ions, having no special attachment to the gelatin, are easily re- 
moved by washing. This shifts the equilibrium to the right, causing more of 
the insoluble salt to go into solution, thus building up the silver-ion concentra- 
tion. Thus more and more silver gelatinate is formed. 

Carroll and Hubbard (Bu. Stand. J. of Res., Vol. 7 (193 1), p. 811) made a 
special study of this reaction. They found that silver gelatinate was formed 
only in the absence of other substances, such as ammonia or sulphites, which 
also form complexes with silver ions. The gelatin-silver complexes are easily 
decomposed by the treatment with either ammonia or sulphite. The fogging 
effect produced when an emulsion is washed for a long time with water prior 
to development, can be traced to this. 

It is very important, therefore, if the last ounce of perfection is to be obtained 
from the reversed screen plates, that the soluble silver ions be completely lack- 
ing in the emulsion after the treatment with acid dichromate. To insure one's 
self of this, it becomes desirable to give the reversed plate a bath in ammoniacal 
sodium sulphite, or 'if the volatile character of ammonia is objectionable, a 
sulphite bath to which ethylene diamine or mono-ethanolamine has been added. 



182 HISTORY OF COLOR PHOTOGRAPHY 

This operation can be carried out in light, so that the plate becomes developable 
at the same time that it is cleared of the objectionable gelatin-silver complex. 

The final blackening of the remaining silver halides can be accomplished in 
many ways. One very simple method is to treat the plate with sodium sulphide 
solutions. This converts the silver halide remaining in the emulsion layer into 
a black silver sulphide. There are many other agents which can be used to 
sulphurize the halide. These are the compounds, containing the SH group, 
in which the hydrogen is in a labile form. Another direct method for convert- 
ing the silver halides into a black substance is to use hydrosulphite, a substance 
that is not to be confounded with sodium hyposulphite, the ordinary hypo of 
the fixing bath. 

However, the method most generally used is to give the remaining silver- 
halide salts a complete exposure to light, then develop them to form metallic 
silver. There may be some disadvantages to this method. The remaining 
grains are very slow, since all the faster grains have been used up to form the 
first image. Their sensitivity is still further decreased by the treatment with 
acid dichromate, which, while dissolving the silver, also etches out from the 
grain surfaces the sensitivity specks that were not previously affected by light. 
There is no telling, therefore, just what exposure is required to make all the 
remaining grains developable. 

Many substances are known which produce intense fog when brought in 
contact with the silverrhalide grain. One such was mentioned above, during 
the discussion of mercury hypersensitization. Thiourea, S = C(NH2)2, is an- 
other chemical that has this property. Indeed, so intense is its action that a 
concentration of one part in a million is sufficient to give considerable fog. 
Many other compounds, all characterized by the — (CSH) group, in which the 
H is a labile hydrogen, will have this effect in alkaline solutions which favor 
the formation of the mercaptan salt. Many dyes also react in this manner. 
Methylene blue, janus green, capri blue, nile blue 2B, cyanine blue, Victoria 
blue, Bindscheller's green, malachite green, methyl violet, and crystal violet are 
all dyes which create considerable fog. Several patents covering the use of 
such fogging agents have been issued to Eastman Kodak Co. and to the I.G., but 
the validity of such disclosures is open to doubt, as the phenomenon is a well 
known one and its application to reversal procedure very natural. 

After the second development is complete, it is desirable to fix the plate to 
remove the last traces of undeveloped silver halides. It is an impossibility to 
develop all the remaining grains, so that some will always remain in the final 
image unless removed by treatment with hypo. Greater clarity will result, since 
the silver salt is a diffusing agent. 

Some of the methods discussed in the chapter on monopacks can be readily 
applied to the processing of screen plates. Thus, after the first image has been 
removed, the remaining silver halides can be converted into silver iodide by 
treatment with the potassium salt of this halide. In this condition, the salt 
will absorb basic dyes such as auramine, rhodamine B, and methylene blue. 



PROCESSING SCREEN PLATES 183 

These can be mixed in such ratios that a black image results. Or a black basic 
dye can be found which is absorbed by the silver iodide. Another possibility 
lies in the use of pinatype black M or platinum black, dyes which stain soft 
gelatin only. The silver image is not removed in this case by treatment with 
acid dichromate, but is converted into silver chloride by treatment with a 
bromoil bleach. This is compounded from dichromate, acid, copper sulphate, 
and a halide, such as sodium chloride or potassium bromide. Besides converting 
the silver into silver chloride or silver bromide, this tans the gelatin in the im- 
mediate vicinity of the negative image, making it no longer receptive of the 
pinatype dyes. The silver halide is then removed by treatment with plain 
hypo, which is without effect upon the gelatin. Treatment with platinum 
black, or pinatype black M, will then stain only the regions outside of the 
negative image, giving a black positive image, whose contrast can be controlled 
to a great extent by the concentration and other physico-chemical properties of 
the solution. 

(C). Photographic Reversal Methods. — These methods are at present 
(1943), merely laboratory curiosities, incapable of yielding images that are com- 
mercially acceptable. They are discussed here, because there is no telling when 
the procedures will emerge from the laboratory stage, and also because it is 
desired to make this discussion of reversal technique representative of the field. 
A plate or film is given a uniform overall exposure sufficient to yield a deep 
black of the desired shadow intensity were development to proceed directly. 
Instead of development, the plate is treated with a solution containing a light- 
sensitive substance whose degradation product is an oxidizing agent, capable 
of destroying the latent image previously formed. The plate is then exposed in 
the camera. Where the light is incident upon it, the latent image formed by 
the first uniform exposure is reduced or destroyed, depending upon the in- 
tensity of the light. Development then yields a direct positive. The applica- 
tion to screen plates is of course obvious. 

The idea can be traced back to Bayard, one of the founders of the Societe 
francaise de Photographic In 1839, several months before the Daguerreotype 
process was made public (Neblette, "Photography, Principles and Practice" 
second edition (1931), p. 18), he exhibited prints that were made in accordance 
with the following technique. Paper, soaked with ammonium chloride, was 
* dried, then floated on top of a silver-nitrate solution. After drying, it was 
exposed to light until a deep black resulted. It was then bathed in a solution of 
potassium iodide, and exposed in the camera while wet. Under the action of 
light, the potassium iodide decomposed to yield free iodine. This united with 
the silver formed in the first printing-out operation, to form white silver iodide. 
Thus where the light had the greatest intensity, complete bleaching of the silver 
took place. A direct positive resulted. 

The photo-chemical decomposition of a latent image is therefore at least as 
old as the idea of photography, and considerably older than any of the photo- 
graphic processes now in use. It is extremely unfortunate that Bayard worked 



184 HISTORY OF COLOR PHOTOGRAPHY 

before developing-out emulsions were known, for then he would have been 
able to apply his reaction to a substance that was thousands of times as fast as 
the printing-out emulsion that he used. He would not have had to expose the 
paper until a suitable final image was obtained, but only until the released 
iodine became sufficient to destroy the developability of the grains in the high- 
lights. The use of potassium iodide as the light-sensitive material would have 
forced him to utilize physical development, as silver iodide is not readily de- 
veloped chemically. But then, there would have been an incentive for the 
photographic technician to devise methods to allow the chemical development 
of silver-iodide emulsions. Other substances whose photo-chemical decompo- 
sition products are oxidizing agents, are the peroxides, the alkali halides, 
dichromate, etc. Perhaps the dichromates could be compounded with a sub- 
stance that yields an acid by photo-chemical decomposition, in which case 
local desensitization would occur by the action of acid dichromate upon the 
latent image. There are many other possibilities, some of which have been 
discussed by J. S. Friedman (Am. PhoL, Vol. 33 (1939), p. 212). 

The use of dichromate for the preparation of direct positives, was proposed 
by G. O't Hooft (Brit. J. PhoL, Vol. 85 (1938), p. 229). A piece of Kodak 
contrasty bromide is bathed in a 2\ per cent solution of potassium dichromate 
to which 10 drops of 10 per cent potassium bromide have been added for every 
10 cc of solution. The excess liquid is removed by wiping, and a thin sheet of 
Kodaloid or other transparent waterproof film is placed over it. The trans- 
parency whose copy is desired is then placed on top of this, and the exposure 
made through it. This will require approximately twenty minutes to daylight. 
The bromide is then washed for one hour to insure the complete removal of the 
dichromate, after which it is given a diffuse overall exposure. Where the light 
acted upon the dichromate, in the previous exposure, a reaction took place 
which released an oxidizing agent. This destroyed the sensitivity of the grain 
in its immediate neighborhood. The exposure through the transparency 
therefore formed an image in the bromide paper that consisted of insensitive 
silver bromide. The remaining silver bromide becomes developable, when ex- 
posed to light, which could be accomplished in the normal manner. 

A variation of this technique would utilize a silver image. The plate or film 
is exposed in the ordinary manner. It is then developed normally, but instead 
of being fixed, it is treated with an acid-dichromate-alkali-halide solution, such 
as the one made by mixing equal portions of A and B. 

Solution A. 

Potassium bromide 30 parts 

Copper sulphate 30 parts 

Hydrochloric acid (1 : 10) 5 parts 

Water to 500 parts 

Solution B. 
Ammonium dichromate 5 parts 

Water to 500 parts 



PROCESSING SCREEN PLATES 185 

The time in this solution should not exceed ten minutes. If the silver is not 
completely bleached, more of solution A should be added. After the image is 
bleached, the plate is rinsed in water, then bathed in dilute (1:200) sulphuric 
acid. This is done to remove the stain. A twenty-minute wash removes all 
undesired soluble substances remaining in the emulsion layer. It is then given 
a diffuse flash exposure to make the silver-halide grains developable. But the 
new grains, which were formed by the action of the bleach upon the negative 
image, are very insensitive and require a considerable amount of light before 
they become developable. Thus a positive image is formed by the final 
development. 

This reaction does not differ materially from one that has time and time 
again been proposed for the formation of a duplicate positive. A film or plate 
containing a silver halide emulsion is further sensitized by bathing in a i\ per 
cent ammonium dichromate solution, and dried. It is then exposed, by con- 
tact, through the negative or positive transparency whose duplicate is desired. 
A thorough wash removes the excess dichromate, after which the plate is given 
an overall diffuse exposure. The grains that had received a previous exposure 
to light while sensitized with dichromate, have been rendered relatively in- 
sensitive and incapable of forming a latent image. Therefore, by the final step 
of normal development, there is formed a silver image which is of the same 
character as the one through which the first exposure had been made. 

In his comments upon the O't Hooft reaction, J. S. Friedman (Am. Phot., 
Vol. 33 (1939), p. 62) pointed out that if the photo-chemical decomposition of 
dichromate results in the release of nascent oxygen, it is possible to utilize the 
reaction in a different manner, to yield a direct positive. The silver halide 
emulsion is first given a diffuse overall exposure, one sufficient to yield a black 
of the desired intensity in the final image were development to proceed im- 
mediately. The latent image thus formed is a mild oxidizing agent, but one 
which is also capable of reacting with still more powerful oxidizing agents. 
This is because the latent image really consists of colloidal specks of silver, 
' adsorbed upon a silver-halide grain. The mild oxidizing properties originate 
in the fact that such a system adsorbs reducing ions such as hydroquinone, 
metol, etc., and in that manner increases the local concentration of these ions 
to such an extent that reduction of the silver-halide grain takes place immedi- 
ately. The oxidizing properties are thus sort of indirect, due only to a favorable 
combination of ingredients. If some substance other than silver halide, which 
was not susceptible to reduction, was the constituent upon which the silver 
speck was adsorbed, then no reduction would be possible. 

The strong reducing action of the latent image toward oxidizing agents of 
even the mildest character, is due to the fact that finely divided silver is very 
easily attacked by these substances, to form silver -ions. One agent that has 
an action too mild even for the latent image, is dichromate that is strictly* 
neutral or slightly on the alkaline side. But this agent becomes pronouncedly 
strong when exposed to light. What is more natural, therefore, than to com- 



186 HISTORY OF COLOR PHOTOGRAPHY 

bine the two? After the diffuse exposure, the plate or film is sensitized with a 
solution of potassium dichromate to which just sufficient ammonia has been 
added to change the orange color to a lemon yellow. After drying, the plate is 
exposed to the negative whose duplicate is desired. It is then washed thor- 
oughly to remove unreacted dichromate, and finally developed in the normal 
manner. A direct duplicate should result. 

The Herschel effect represents probably the most interesting of all of these 
theoretical processes for the formation of a direct positive. By far the greatest 
contributor to this field has been Luppo-Cramer, and a summary of his con- 
tributions .is contained in the latest edition of Eder's "Handbuch der Photo- 
graphic" (Vol. 2, part i, Grundlagen der photographische Negativeverfahren; 
Vol. 3, part 3, Sensibilisierung und Desensibilisierung) . Carroll and Kretchman 
(Bu. Stand, J. of Res., Vol. 10 (1933), p. 499)> K. Weber {Phot. Korr., Vol. 
75 (i939)> P- 22 ) an d J. S. Friedman (Am. Phot., Vol. 33 (1939), p. 848) have 
also disclosed interesting studies of this phenomenon. 

When a plate is exposed to form a latent image, the exposed silver-halide 
grain takes on an added absorption in the red and infrared. If the latent image 
is exposed to light in this range, it becomes reduced in intensity, and may even 
be completely destroyed. Luppo-Cramer and many other experimenters have 
utilized this result to prepare positives directly. K. Weber proposed the 
following routine: A sheet of gaslight paper is exposed diffusely for three 
seconds to the light of a 60-watt bulb at a distance of 50 centimeters. It is then 
bathed for three minutes in a solution that contains one gram of pinakryptol 
yellow in two liters of water, and finally dried. A five-hour exposure in the 
camera through a red filter gave a direct positive, whose characteristics were 
typical of infrared emulsions. 

But if an Ansco Brovira or other paper of equal speed is used, with a slightly 
different technique, a three-minute exposure becomes sufficient. The paper is 
first diffusedly exposed as above. It is then bathed in Renwick's solution for 
10 minutes. 



Potassium iodide 


10 parts 


Sodium sulphite 


10 parts 


Hypo 


30 parts 


Water to 


1000 parts 



This is followed by a half-hour wash in running water, then by sensitization in 
a solution made as follows: 

Malachite green 0.04 part 

Potassium bromide 10.0 parts 

Water to 1000 parts 

The time of sensitization is 3 minutes at 65 F, followed by rapid drying. The 
exposed paper can be developed in alkaline amidol, which must be mixed just 
before use. 



PROCESSING SCREEN PLATES 187 



Solution A. 




Amidol 


5 parts 


Sodium sulphite 


50 parts 


Water to 


500 parts 


Solution B. 




Sodium carbonate 


50 parts 


Water to 


500 parts 



For use, mix equal quantities of A and B. It was Luppo-Cramer who dis- 
covered that the addition of bromide to the dye solution gave a greatly en- 
hanced effect. Carroll and Kretchman proved that the effect is sensitized 
by desensitizing dyes. In their technique, test plates of a special silver bromide 
emulsion were first given an overall diffuse exposure to white light. They were 
then bathed in the dye solution with and without bromide, the excess dye re- 
moved by wiping, and finally dried. The exposures were made in a spectro- 
graph, to determine the color sensitivity of the result. The final development 
was with D-11, a process developer. In particular, one experiment is very 
interesting. A process plate was given uniform exposure to white light, then 
bathed in a solution that contained 2-w-nitro-styrrylquinoline dimethyl- 
sulphate 1 : 5000, and potassium bromide. After drying, this was exposed for 
2\ minutes under an Eder-Hecht wedge, at a distance of 30 centimeters from 
a 95-watt gas-filled lamp. Development was with D-n. 

The effect of bathing with dyes is best illustrated by the following experiment, 
made by Carroll and Kretchman. A pure silver bromide emulsion was pre- 
exposed for several minutes, then bathed in a one per cent solution of potassium 
bromide and no dye. After drying, this was exposed in the spectrograph 
through a red filter, to test the pure Herschel effect. It required a 150-minute 
exposure to produce even a limited reversal. But when dyes were added to the 
potassium bromide solution, very good reversals were obtained in ten minutes. 
The opening paragraph of the Carroll and Kretchman paper states that positive 
images of passable gradation may be produced by this process, the sensitivity 
being about one-tenth that of a lantern slide. 

(D). Lumiere Filmcolor. — Of the four screen-plate processes (Lumiere 
Filmcolor, Agfacolor, the Finlay Plate, and Dufaycolor) formerly available to 
the technician, Dufaycolor is the only one which at the present writing (1944) 
remains commercially available, but the methods of processing all these mate- 
rials, which were carefully worked out by the manufacturers, are given for their 
historical value. 

Lumiere Filmcolor, also known as Lumicolor, is supplied only in the form 
of film. The speed of this material is one-twelfth normal, whereas the previous 
material had a speed that was one-sixtieth normal. 

After exposure, the film is developed for 2§ minutes in a special solution 
whose composition is as follows: 



188 HISTORY OF COLOR PHOTOGRAPHY 



Metoquinone 


15 parts 


Sodium sulphite 


100 parts 


Ammonia (22 Be) 


32 parts 


Potassium bromide 


16 parts 


Water to 


1000 parts 



For use, dilute with 4 parts of water. Metoquinone is a molecular combination 
of metol and hydroquinone, and is sometimes marketed under the trade name 
Quinomet. If desired, the exposed film could first be desensitized by treatment 
with a phenosafranine or pinakryptol green solution, then developed by in- 
spection. If this be done, it is recommended that development be carried to a 
Watkins factor of 10. This means that the time from the start of development 
to the first appearance of the image (disregarding sky or other extreme high- 
lights) is to be multiplied by ten. Since the recommended time is 2J minutes, 
a correctly exposed film will have image appearance in fifteen seconds. 

It is to be noted that the Lumieres prefer to use ammonia as the silver-halide 
solvent. This necessarily has an influence upon the length of time of devel- 
opment, since the volatility of the ammonia would make the solution quite 
unstable. This can be overcome by the use of the non-volatile ammonia 
substitutes such as ethylene diamine, or mono-ethanolamine. Dr. Troland 
overcame the defect by the use of caustic soda and ammonium halide. The 
two would react to yield free ammonia, but if the caustic be present in a 
molar concentration greater than that of the halide, the loss of ammonia will 
have but little effect upon the alkalinity of the developer. 

After a minute wash, the film is placed in the acid dichromate reversal 
solution. 

Potassium dichromate 2 parts 

Sulphuric acid, concentrated 10 parts 

Water to 1000 parts 

In making up the solution, the dichromate is first dissolved in water, then the 
acid added cautiously. It must be carefully noted that the acid is to be meas- 
ured out in a dry graduate, and added to the water, never the reverse. When a 
little water is added to sulphuric acid, tremendous heat is evolved, and there 
is great danger that the water will begin to boil and spatter this harmful agent 
indiscriminately. If the solution is warmer than 65 F it should be cooled to 
that temperature before the film is placed in it. Four minutes is sufficient time 
for the silver to be completely removed, but there is no harm in letting it re- 
main a longer time to insure this, as otherwise streaks will occur. At this point, 
if the film is examined, the highlights would be just barely veiled over. If a 
heavy deposit remains, it is evidence of underexposure. After the film has 
been in this solution for a minute, white light can be turned on, and the rest of 
the processing done in bright light. 

A rinse in water is followed by a second development which, in cold weather, 
can be done in the same developer used initially. But in warm weather, when 



PROCESSING SCREEN PLATES 189 

the temperature of the washes can rise well above 70 F, the use of a high-alka- 
line solution creates the danger that the emulsion may dissolve away, a defect 
known as frilling. If the second development be carried out in an amidol de- 
veloper such as: 

Sodium sulphite 30 parts 

Amidol 5 parts 

Water to 1000 parts 

the danger of excessive swelling is eliminated. In this solution, the develop- 
ment is carried out to completion. A wash, and the drying operation, com- 
plete the procedure. 

If the original exposure has been too short, a heavy, dull transparency will 
result which it is very difficult to correct by reduction, since there will be not 
only blocked highlights but also a predominant blue hue, which is due to the 
fact that an excess of blue elements is present in the screen to overcome the 
limited blue component present in ordinary "white" light. 

Overexposure will yield very transparent films that are weak and lacking in 
contrast. If the color of the highlights is not too warm, this defect can be 
overcome by an intensifying bath, whose composition is: 

Sodium sulphite 10 parts 

Mercuric iodide 1 part 

Water to 1000 parts 

The treatment is continued until the desired contrast is obtained, after which 
the film is washed, redeveloped in amidol for five minutes, washed again, and 
finally dried. The intensified image will not keep unless the redevelopment is 
carried out. The transparency and brilliance of the colors can be enhanced 
by a coat of varnish. 

(E). Agfacolor. — The Agfacolor film is a screen material that is practi- 
cally identical with Filmcolor in structure, but is slightly faster. The proc- 
essing is also quite similar to the other. The preferred developer has the 
composition 

Metol 3.24 parts 

Sodium sulphite 25 parts 

Hydroquinone 1 part 

Potassium bromide 1.5 parts 

Ammonia (0.91) 7.5 parts 

Water to 1000 parts 

Use full strength. Develop three minutes at 65 F for the plate, and four minutes 
for the film. As with Filmcolor, Agfacolor can be desensitized by treatment 
with a 1 : 5000 pinakryptol green solution, for two minutes. In that case, the 
development can be carried out in a bright red light, to a Watkins factor of 
ten. The treatment beyond the first developer stage is identical with the one 



i go HISTORY OF COLOR PHOTOGRAPHY 

disclosed for Filmcolor, and the processing baths differ from those used in the 
first only in minor details. 

Reversal bath: 

Potassium dichromate 50 parts 

Sulphuric acid, concentrated 108.5 parts 

Water to 1000 parts 

For use add 10 parts of water to one part of stock. 

Second Developer: 

Sodium sulphite 57 parts 

Amidol u§ parts 

Water to 570 parts 

If the first developer be used for second development, it becomes unfit for 
further use, due to the carrying over of some reversal solution. But then, 
developers compounded with ammonia as the activating agents cannot be 
used too long, because of the volatility of the ammonia. 

(F). Finlay Plate. — We will leave until last the discussion of Dufaycolor 
processing, and take up at this time the Finlay plate. This is an entirely 
different type from the others in that the screen is separated entirely from 
the photographic emulsion. The screen is coated upon a glass plate, which is 
then used as the front element of the bipack, the rear element of which is a 
special panchromatic emulsion, prepared either by the Ilford Company 
(England), or the Eastman Kodak Company. The factor occasioned by the 
use of the screen, is six. Up to the advent of the Dufaycolor film, this was the 
most rapid screen plate made. The results obtained by it were equally as good 
as those from Autochromes and Agfacolor material. 

The separation of screen and emulsion enabled easy duplication of the trans- 
parency. Color-separation positives could be made directly from the original 
negative by use of "block-out-screens," which is a sort of masking procedure. 
Since the advent of the fine-grained and exceedingly fast emulsions, the Finlay 
plate should be able to give speeds that are far beyond the present scope of 
Dufaycolor or Kodachrome, but evidently very little is being done along these 
lines. Perhaps the use of a screen that has but 175 elements to the inch is too 
coarse. But where no enlargement is intended, the scheme offers untold 
possibilities. Instead of utilizing inefficient filters to make possible the use of 
the same plate under different lighting conditions, it becomes possible to use 
different emulsions, each made for use for a special type of illumination. 
Since the first appearance of the Finlay plate it has become possible to prepare 
emulsions with practically any desired characteristics. 

The bottom of the Finlay plate screen carries a series of color-registration 
marks. If this area constitutes a highlight, the registration marks will be 
automatically registered upon the negative. If it constitutes a shadow, it is 
important to give the area a supplementary exposure to white light. This 
can be done by replacing the normal slide with one which has the correspond- 



PROCESSING SCREEN PLATES 191 

ing section cut away, then exposing (through the camera), to the light reflected 
from a sheet of white paper. The registration marks in the Finlay screen 
consist of bands of red and blue, separated by green circles. When the positive 
is registered on top of the viewing screen, these should appear in the comple- 
mentary colors, that is as cyan and yellow bands, separated by magenta 
circles. Most of the texts give these colors as blue or green, yellow, and red 
respectively, but where this is done, it represents a lack of preciseness on the 
part of the writers. 

The original Finlay negatives are developed to yield the maximum coverage 
of the density range of the original. This means that a relatively low contrast 
is desired. Three types of developers are recommended. 

D-76 

Metol 2 parts 

Sodium sulphite 100 parts 

Hydroquinone 5 parts 

Borax 2 parts 

Water to 1000 parts 

Develop 12 to 14 minutes at 65 F. 

Glycin Developer 

Sodium sulphite 37 J parts 

Glycin 15 parts 

Sodium carbonate monohydrate 75 parts 

Water to 1000 parts 

For use dilute with an equal part of water. Develop 5 minutes at 65 F. 

Metol Developer 

Solution A. 

Metol 12 J parts 

Sodium sulphite 50 parts 

Water to 1000 parts 

Solution B. 
Potassium carbonate 62J parts 

Water to 1000 parts 

For use take one part of A, one part of B, and 5 parts of water. Develop for 
6 minutes at 65 F. This is the temperature best suited for materials to be used 
with the Finlay process. At higher temperatures the gelatin will swell too 
much and after fixation there will be danger that the shrinkage will not be 
back to normal, thus throwing the image out of registry with the screen ele- 
ments. The fixing bath should be compounded without alum, and should 
be used once and then discarded. 

The negatives, when examined closely, will show the image to consist of 
minute squares of silver, varying in density. These correspond identically to 
the areas of the colored screen elements. No enlargement or reduction of the 



192 HISTORY OF COLOR PHOTOGRAPHY 

positives made from a Finlay negative is possible, as then the image squares 
would no longer correspond to the screen. Therefore, the positives are to be 
made by contact only. These should be brilliant and clear. " Eastman Process 
or the equivalent Hammer plate are recommended. Ilford makes a special 
material, known as the Finlay-Ilford positive plate, for this purpose. 
After exposure, the positive transparency plates are developed in D-72. 



D-72 Stock solution. 




Metol 


3 parts 


Sodium sulphite 


45 parts 


Hydroquinone 


12.2 parts 


Sodium carbonate 


67.5 parts 


Potassium bromide 


1.9 parts 


Water to 


1000 parts 



For use add four parts of water to one part of stock. If the negative is exces- 
sively flat, a hydroquinone-caustic developer should be used. 



Solution A. 


• 


Hydroquinone 


25 parts 


Metabisulphite 


25 parts 


Potassium bromide 


25 parts 


Water to 


1000 parts 


Solution B. 




Potassium hydroxide 


50 parts 


• Water to 


1000 parts 



For use mix equal parts of A and J5. Develop 90 seconds at 65 F. 

The final colored transparency is made by registering the positive with a 
special viewing screen. These can be obtained in quantity, thus enabling the 
operator to make as many duplicates as are necessary. 

Instead of making the positives upon process material, it is possible to make 
them upon a special positive emulsion that is coated on top of a viewing screen. 
When this is done, the positive and negative must be registered in the dark- 
room. The screened positive plate contains an extremely slow blue-sensitive 
emulsion, one which can safely be manipulated in the light of a Series OA 
safelight containing a 50-watt bulb. The two plates are shifted until the regis- 
tration marks appear as a green outer band, a yellow inner band, and a red 
circle. It must be remembered that a minus-blue light is being used, so that 
the magenta will appear as red, and the cyan as green. Each patch must be 
completely devoid of the complementary color when viewed under a ten- 
power magnifier. When this is accomplished, the two are clamped together 
tightly, and exposed through the negative side. The color screen on the posi- 
tive lies between the emulsion and the glass base, hence the exposing light does 
not go through this until after the emulsion has received a latent image. Then 
it acts as an anti-halation backing. There is no particular advantage in the 
use of a screened positive emulsion other than the possibility that loss of 



PROCESSING SCREEN PLATES 193 

registry may result from faulty processing, due to an uneven expansion and 
contraction of the emulsion on the glass plate. But if ^here is any difficulty in 
obtaining registry, and a very long time is required before this is accomplished, 
there is grave danger of fog setting in, giving dull and lifeless prints. 

(G). Dufaycolor. — This is by far the' most popular of all the screen proc- 
esses. It has a speed rating which is approximately the same as Kodachrome. 
While at the present writing, Kodachrome can be processed only in central 
laboratories, Dufaycolor processing can be accomplished immediately, giving 
the operator the opportunity to make immediate checks. 

Dufaycolor can be processed in two ways. It can be reversed to yield a direct 
positive, from which separations can be made if paper prints or other type of 
duplicates are desired. It can also be developed as a negative that is com- 
plementary in color as well as in tone. If copies are to be made on Dufaycolor 
material, the second is the preferred technique. G. B. Harrison and D. A. 
Spencer (Phot. /., Vol. 77 (1937) April) made a detailed study of this, detailing 
the problems that are involved. 

The accurate rendition of screen processes depends upon the correct repro- 
duction of densities behind each element of the mosaic. If a color-screen 
negative be projected upon another screen plate, one error arises due to the 
overlapping transmissions of the screen elements. Since it is impossible to 
accurately register the green elements in the master with the green elements 
in the copy, the projection of some of the green elements will fall on red and 
blue elements in the copy. If the three colors are mutually exclusive, no harm 
other than a loss in contrast (which can be overcome by a longer development) 
will result. No deposition of green densities in areas that are reserved ex- 
clusively for red and blue, will take place. But the elements in Dufaycolor 
are not mutually exclusive. The greens transmit considerable blue and red. 
Therefore, by such projection, green densities will register behind red and blue 
areas, thus diluting and desaturating the greens. The same is true for the 
other two elements, although to a smaller extent. This, incidentally, is the 
one serious objection to Dufaycolor, but one which can be rectified by a pro- 
cedure known as masking. 

Another cause for poor results when making copies from the screen plates on 
screen material, is due to scatter of the exposing light by the silver-halide grain 
in the copy material. This type of error was discussed in an article published 
in The Photographic Journal (Vol. 73 (1933), p. 19). It was shown that the 
densities behind a given dot would spread a little due to this scatter, and yield 
densities behind adjacent dots of different colors. Two effects would be 
produced. First the opacity behind a desired dot would be reduced, since 
some of the exposing light is not used; and second, the opacities behind 
adjacent dots would be increased. The net effect is a loss in color contrast, 
brilliance, and saturation. If developers are compounded with ammonia or 
other silver halide solvents present, a peculiar effect is obtained. A reduction 
of density occurs immediately adjacent to a heavy-density region. This is 



194 HISTORY OF COLOR PHOTOGRAPHY 

directly in opposition to the effect of scatter, so that the one neutralizes the 
other. In reversal procedure, the use of ammonia or thiocyanate (in amounts 
sufficient to obtain neutralization of the effects of scatter) is permissible, since 
the resultant high fog level is of no consequence. But where the film is to be 
developed as a negative, the presence of excess fog ruins the printing qualities, 
giving dull and lifeless prints, A substitute had to be found, therefore, for 
ammonia or thiocyanate. The *f ollowing processing techniques were tried: 

i. The film was developed in a very weak solution. In this manner, the 
grains closest to the screen elements developed first, since they were exposed 
first, and received the maximum intensity of exposure. The development of 
these grains used up the developer, so that the upper strata of emulsion grains 
could develop only after the lower ones were completely reduced. If the de- 
velopment was stopped short of this, the upper emulsion grains (representing 
the portions in which scatter is at a maximum) remained undeveloped. A very 
weak image resulted, but this could be intensified. However, the use of very 
weak developer solutions gave rise to variations, loss of speed, and other defects 
that forced the abandonment of the technique. 

2. The negative is considerably overdeveloped and the image in the upper 
portions of the emulsion layer is removed by treatment with Farmer's reducer 
to which is added a considerable concentration of alcohol (up to 70 per cent), 
sodium sulphate (30 per cent), glycerin, or other substances which reduce 
diffusion. The difficulty in the accurate control of the diffusion forced the 
abandonment of this procedure. 

3. The latent image in the upper strata of the emulsion layer was destroyed 
by treatment with oxidizing agents, prior to development. 

4. A surface image is developed, then removed by treatment with acid 
permanganate, the film washed, and finally developed normally. 

5. A surface image is developed with a non- tanning developer, and then the 
rest of the development is continued with a tanning developer. When this is 
treated with Farmer's reducer, the tanned gelatin will prevent the action in 

, the lower portions. Thus only the surface image will be removed. 

For proper results, processes 3 and 5 all depend upon controlled diffusion. 
With the use of thinly coated emulsion layers, this is faulty and unreliable. 

6. The latent image is developed in a solution whose action is slowed down 
by the products of the development. Since the action starts at the bottom, 
next to the screen, this will give preferential treatment to this part of the 
image. A developer to accomplish this result could be concocted as follows: 

Hydroquinone 12 J parts 

Potassium metabisulphite 125 parts 

Potassium bromide 12^ parts 

Potassium hydroxide 25 parts 

Water to 1000 parts 

This gave good results, but the contrast was too high for good printing quality. 

7. The latent image is developed in a "depth" developer, whose composition 



PROCESSING SCREEN PLATES 195 

is altered as it diffuses through the gelatin. Consider a developer whose com- 
position is as follows: 



Amidol 


15 parts 


Sodium sulphite 


125 parts 


Potassium bisulphite 


50 parts 


Water to 


1000 parts 



As this solution diffuses into the gelatin layer, the bisulphite becomes absorbed 
by the gelatin. Therefore its concentration becomes diminished as the solution 
penetrates to the bottom. Thus there is no restraining action at the bottom, 
and development proceeds there in the normal manner. But the top strata 
are quite acid, so little action takes place there. This is why an acid amidol 
developer acts first at the bottom. 

A developer compounded with hypo in the bath acts in a similar manner. 
The hypo is used up in the upper grain strata of the emulsion layer, so that the 
lower portions are not affected, and develop up in the normal manner. A very 
effective developer of this type for use with Dufaycolor film is: 



Metol 


10 parts 


Sodium sulphite 


30 parts 


Sodium hydroxide 


10 parts 


Hypo 


20 parts 


Water to 


1000 parts 



The use of this developer gave a negative image that lay substantially close to 
the screen elements in the lower portions of the emulsion layer. Potassium 
iodide, and potassium or ammonium thiocyanate, if present in sufficient 
quantity, also give the same result. From negatives prepared in this manner, 
it was possible to print on to screen material after due precautions were taken 
to filter the printing light to correct for the overlaps in the transmissions of 
the screen elements. 

The more normal procedure is to process Dufaycolor by reversal. The 
recommended technique differs but slightly from that described above for use 
with Lumicolor and with Agfacolor. But the differences are all in the direction 
of improved technique. The initial development is carried out in a developer 
compounded to contain potassium thiocyanate, thus: 



Metol 


6\ parts 


Sodium sulphite 


50 parts 


Hydroquinone 


2 parts 


Sodium carbonate 


37 parts 


Potassium bromide 


2 J parts 


Potassium thiocyanate 


9 parts 


Water to 


1000 parts 



Develop for three minutes at 65 F. If the film is desensitized, it can be de- 
veloped by inspection to a Watkins factor of six. 



196 HISTORY OF COLOR PHOTOGRAPHY 

After development, the film is washed in running water for a minute, then 
reversed in acid permanganate. 

Potassium permanganate 3 parts 

Water 1000 parts 

Sulphuric acid, concentrated 10 parts 

The film should remain here until the silver is completely removed, after which 
it is washed again for two minutes, and cleared in a bisulphite bath (2! per 
cent). This serves not only to remove the reduced permanganate, but also 
to break up the gelatin-silver ion complex. At this point the film can be given 
an exposure to white light to make all the remaining grains developable. 
Reduction of the silver salts can be accomplished by treatment in the fol- 
lowing solution: 



Metol 


1 part 


Hydroquinone 


5 parts 


Sodium sulphite 


50 parts 


Sodium carbonate 


20 parts 


Potassium bromide 


1 part 


Water to 


1000 parts 



It is also possible to use the first developer at this point, but in that case it can 
no longer be used with freshly exposed film. . The silver halide grains that escape 
reduction at this stage can be removed by treatment with an ordinary chrome- 
alum hypo bath, or one compounded as follows : 

Solution A. 

Hypo 360 parts 

Potassium metabisulphite 12 parts 

Water 1000 parts 

Solution B. 

Chrome alum 10 parts 

Water 1000 parts 

Add Solution B to A. 



CHAPTER 15 
SEPARATION NEGATIVES 



U. 



P to now we have discussed the problem of making an original color shot 
from the point of view of fundamentals, and apparatus required. Very little 
was said about the actual act itself. To photograph an object in color, the 
operator has at his command a variety of methods. He can use a one-shot 
camera in which the lens beam is divided into three parts by any one of the 
many mechanisms described in previous chapters. In these cases he will 
have as his immediate product three negatives, each of which represents the 
intensities of a single primary as it is present at each point in the original. If 
this procedure represents limitations beyond his control, he can use one of 
the integral processes, such as Kodachrome, Ansco Color, Dufaycolor, etc. 
In these procedures he will have as his immediate end result a colored positive 
transparency, which he can use as a new original from which separations can 
be made. 

Another possibility at his command is to photograph the subject through 
the three niters in succession. This represents the simplest of the methods, 
and the best, but unfortunately it can be used only in cases where the object 
remains perfectly immobile for the time required to make three photographs. 
This is a time long enough to enable one to expose three times, to change 
plates or films three times, and to change filters on a like number of occasions. 
Mechanisms have been devised which do these operations automatically and 
mechanically. These are called repeating backs. A special plateholder is 
used in which the three plates are loaded side-by-side. The three filters are 
placed so that each plate is situated immediately behind one filter. By a 
single operation, the plateholder is moved so that one filter-plate combination 
is in position, the lens shutter is opened for the required time, and it is then 
closed; the plateholder is then moved so that a new filter-plate combination 
is brought into position, and the cycle is repeated until three exposures are 
made. It has been claimed that the entire act can be done in a period of a 
few seconds. The Autotype Company of England, makers of carbro and 
carbon tissue, market several types of such backs that can be used with the 
ordinary plate camera. The repeating back has a slight advantage over the 
normal one-shot camera in that it can be used under any type of lighting, for 
it is possible to vary to any desired extent the exposure behind any filter. 

White light is largely a misnomer, for the normal Mazda light is very much 
on the red side, and carbon arcs are very much on the blue. Consider for 

197 



198 HISTORY OF COLOR PHOTOGRAPHY 

instance, lights with color temperatures of 3000 K, 3200 K, 3435 K and 5400 K. 
The relative distribution of three primaries in these light sources is as follows: 



Degrees K 


Red% 


Green % 


Blue % 


3000 


54 


32 


14 


3200 


49 


34 


17 


3435 


46 


35 


19 


5400 


33 


34 


33 



It is seen that as the color temperature of the light rises, the relative proportion 
of the red rays drops and that of the blue increases, until at 5400 K, which 
is the standard adopted for outdoor work, the ratios are practically identical. 
Ordinary Mazda light operates at a color temperature of 3000 K. Overvoltage 
lamps operate either at 3200 or at 3435 K. It becomes important to know, 
therefore, exactly what type of light is being used, since different light sources 
have different proportions of the three primaries present, and therefore will 
have different values for the filter factors. 

In one-shot cameras, the filter factors are taken care of in the division of 
the intensities among the three sub-beams. Consider a camera designed to 
be used with Mazda light, and an emulsion which has filter factors of 2, 6, 
and 10 for the red, green, and blue filters respectively. In order to obtain 
balanced exposures, it will be necessary to direct five times as much light to 
the blue, and three times as much to the green, as to the red filter. Therefore 
the mirror reflectors are so arranged that 11 per cent of the lens beam is 
directed to the red, ^ per cent to the green, and 56 per cent to the blue filter. 
But if this camera is to be used for light of daylight quality, where the filter 
factors are 4, 6, and 4 respectively, the blue filter image will be considerably 
overexposed, and the red filter image considerably underexposed. Owners of 
one-shot cameras must be sure to use them only under the conditions for 
which they were designed. It is possible to put filters over the lens which will 
convert daylight into Mazda, and vice versa, but this results in a loss of light, 
and one-shot cameras cannot afford such loss. Some cameras can be converted 
easily from daylight to Mazda by merely shifting the position of the filters. 
This means that constant vigilance must be exercised to keep in mind what 
light is being used. However, this is no hardship, for color photography 
should not be attempted unless constant supervision is being exercised. 
With the repeating back, the relative exposure given to any one filter can be 
varied at will. But this can be used only on those occasions when the object 
can be held stationary for at least ten seconds. 

The relative exposures that have to be given through the different filters 
are determined by the values of the filter factors for the given emulsion and 
the given light source. Thus if the factors are 2, 6, and 10, it is necessary to 
expose ten times as long behind the blue, six times as long behind the green, 
and twice as long behind the red filter, as would be required were no filter 
used. 



SEPARATION NEGATIVES 199 

In making exposures, one other thing that must be considered is the failure 
of the reciprocity law. Theoretically the silver density that results from a 
one-second exposure at a light intensity of 100 units, should be the same as 
the density obtained from a 100 second exposure at a light intensity of one 
unit. The product of /, the light intensity, and T, the time of exposure, 
should be a constant. A variation of a certain percentage in one, accompanied 
by the same variation but in the opposite direction, in the other, should give 
the same result. But for some reason this is not so. The density obtained 
when the light intensity is high, is greater than when the exposure time is 
high. This makes itself felt in cases where time exposures are required. 

Most speed ratings are made under the conditions for which the given 
emulsion is intended. For that reason a standard development in the de- 
veloper solution recommended is also desirable, although use of the borax 
developer, D-76, is highly recommended for all cases where a moderate en- 
largement is required. For instance, a given emulsion is rated at Weston 100, 
and it is recommended that this emulsion be developed for fifteen minutes in 
D-76 at 65 F. This implies that when the light intensity is such that instan- 
taneous exposures are being made, not over one second in duration, and the 
film or plate is developed as specified, exposure at a rating of Weston 10a 
would be correct. But if the light intensity is very low, or if the exposures 
are made through dense filters or at a low diaphragm opening so that from 
two to five minutes exposure is required, it will be found that the emulsion 
must be rated at Weston 25 or 35, rather than 100. 

Of course, under the conditions specified above, it is quite immaterial 
whether the speed is 100 or 10, since speed is of no consideration. And it 
may be argued that it is ridiculous to use a fast negative emulsion, that is 
relatively coarse grained, for cases where a moderate-speed emulsion, prepared 
to be used under just these conditions, would suffice. But the object to be 
photographed may have a density range of 2.3. It is only when using the 
fastest known materials that this range may be successfully copied, so the 
use of such high-speed emulsions may be essential to the subject matter at 
hand. Then again, it may be desired to use Kodachrome film to copy a series 
of paintings or to make some still setups. Kodachrome can be obtained only 
in two grades, outdoor and indoor. There are no fast and slow Kodachrome 
materials, one suitable for instantaneous exposures, the other for long ex- 
posures. The same material must be used for both purposes. In such cases, 
the speed of the Kodachrome should be rated at a maximum of one-half 
normal when time exposures are being made. It is to be hoped that the rec- 
iprocity law failure is uniform for the three primary colors. 

In the processing of the separations, it is desirable that the contrast ranges 
be identical in all three negatives. For this reason the inclusion of a gray scale 
is of the utmost importance. The densities in the gray scale should be at 
least equivalent to the densities encompassed in the subject matter itself, and 
preferably should overlap them. Then if the exposure and development 



200 HISTORY OF COLOR PHOTOGRAPHY 

be so arranged that all the steps in the gray scale lie on the straight-line portion 
of the curve, it is certain that this same absolute proportionality will exist 
with respect to the densities in the original subject. 

When such a gray scale is present the operator can completely forget the 
subject, and concentrate his attention upon the scale. If this be identical in 
the three negatives, and if the different steps of the scale all lie on the straight- 
line portion of the H & D curve, then the three negatives are completely 
in balance and are correctly exposed. The existence of a toe would indicate 
underexposure. The existence of a shoulder would indicate overexposure. 
If the scales show both toe and shoulder, this is indicative that the density 
range in the gray scale is too great for the particular emulsion, and another 
type of material should be used. 

The latitude of an emulsion may be considerably increased by a reduction 
in contrast. By this is meant that if the latitude is three octaves at a contrast 
of 1.20, it would possibly have a latitude of four or maybe five octaves at a 
contrast of 0.80, and six or more octaves at a contrast of 0.60. Three octaves 
would correspond to a density differential of 0.90. At a contrast of 1.20, the 
density differential would be 1.20 X .90 or 1.08. This means that at this 
high gamma, the differential in densities between the two extremities of the 
straight-line portion of the curve would be 1.08. The shoulder would be 
approximately .10 density unit above the upper value, and the toe approxi- 
mately .20 below the lower value, so that the density differential in the nega- 
tive would be 1.08 + 0.10 + 0.20, or 1.38. Six octaves correspond to a 
density differential of 1.80, but at a contrast of 0.60 the actual density dif- 
ferential in the straight-line portion of the curve would be 1.80 X 0.60, or 
1.08. The toe and shoulder would add approximately 0.30 density unit, 
so that in this case, the negative would have a density differential of 1.38, 
but this differential correctly images six octaves. The same density differential 
in the case of a more contrasty negative images correctly only three octaves. 

The filter factors of panchromatic emulsions indicate that the red sensitivity 
of the emulsion is by far the greatest, the green next, and the blue least. A 
typical set of factors would be 2, 6, and 10, for the red, green and blue sensitivi- 
ties. But the light for which these factors hold has a color temperature of 
approximately 2800 K. It contains 60 per cent of its total intensity in the 
red, and only 13 per cent in the blue. It is a well-established fact that the 
sensitivity of the emulsion is predominantly in the blue. If the emulsion were 
exposed to equal quantities of red, green and blue lights, then it would be 
seen that well over 70 per cent of the total sensitivity of the emulsion lies 
in the blue. Therefore, the emulsion is a much better absorber of blue light 
than of red or green. This means that the blue image is much more of a 
surface image, hence it is flatter than the other two. To achieve the same 
contrast through the C5 filter as through the A or B, it becomes necessary 
to develop this negative for a greater length of time, from 25 to 50 per cent 
longer than the other two. Since a longer time of development means that 



SEPARATION NEGATIVES 201 

the H & D curve becomes shifted somewhat to the left, a slight increase in 
emulsion speeds results so that a slightly shorter time of exposure could be 
used. The exact amount of adjustment must be determined for each emulsion, 
but once determined, remains quite constant. 

Ansco and the Eastman Kodak Company have set a good example to the 
rest of the industry by the publication of a pamphlet in which the physical 
and photographic characteristics of all their film materials are described. 
Not only are the normal properties listed, but there are given wedge spectro- 
grams (both to daylight and to tungsten) ; filter factors for the more popular 
filters (which also include the A, B, and C5 set); exposure tables (including 
Photoflood and Photoflash data); recommended developer times for several 
developers; time-gamma, and time-temperature curves for these developers; 
and a set of H & D curves for one of the developers, giving the complete pic- 
ture for values of gamma ranging from 0.50 to 1.00. One other very important 
bit of information that is given, is the resolving power of the emulsion. Thus 
the Super Panchromatic 35mm film is given a resolving power of 45 lines per 
mm with D-76. If the images of a series of parallel lines lie 45 to the milli- 
meter, they will be distinct, but if they are more than 45 lines to the milli- 
meter, they will not appear as individual lines, but as an indistinct gray area. 

So much for the preparation of separations by the direct processes. Auto- 
chrome, Dufaycolor, and the other integral processes, give for the immediate 
products a single image in which the three color densities lie next to each 
other. To make reproductions, it is necessary to separate the three upon 
separate emulsions. The different integral processes allow this to be done in 
a variety of ways. Kodachrome, Ansco Color, Agfacolor Neue have the three 
densities lying one over the other, differing from each other in their respective 
spectral absorptions and transmissions. Therefore, the different strata may 
be copied by the use of colored lights. 

The Kodachrome picture consists of a yellow image that lies in the top emul- 
sion layer, and which depicts the blue densities. Below this, in a central 
layer, lies a magenta image depicting the green densities. In a bottom layer, 
lying adjacent to the celluloid, there is a cyan image giving the red densities. 
Light that is transmitted through a film of this type must traverse all three 
layers. If this be white light, then by its passage through a given point in the 
top layer the yellow image at that point will abstract a sufficient quantity of 
blue light to yield the requisite blue density. Such yellow dyes transmit 
(theoretically at least) the green and the red rays completely, so that no green 
nor red will be removed from the light source at this point. The remaining 
blue rays will pass unhindered (theoretically) through the magenta and cyan 
images. Hence the yellow image will form, or cause to be formed, only the 
blue densities. The rest of the light will now proceed to the magenta layer. 
Here just sufficient green light will be subtracted to yield the desired green 
density, the blue and red portions going through undisturbed. In a similar 
manner, by its passage through the final cyan layer, red densities will be 



202 HISTORY OF COLOR PHOTOGRAPHY 

formed to the proper degree. The light that reaches the opposite side will 
contain red, green, and blue densities, that, in combination, reconstruct the 
color as it was present at the corresponding point of the original. 

Now consider what happens if instead of using white light, there is used the 
light that passes through a C4 or C5 filter. This corresponds to the blue pri- 
mary. Upon passage through the yellow layer, the intensities of the blue light 
will be varied at every point to correspond to the densities of the blue primary 
present at each point. We say that the yellow image modulates the blue 
light passing through it. No further action upon this light will take place 
upon its passage through the other two layers since they are both transmitters 
of the blue. Therefore, if this ligjit is made to fall upon a photographic emul- 
sion, it will register only the blue densities that are present in the Kodachrome. 
In this way a blue-filter separation is obtained. Similarly, if the color trans- 
parency is illuminated by the light that is passed through the green filter, 
it becomes possible to obtain the green-filter separation. By the use of red 
light, it is possible to obtain the red-filter separation. In other words, it 
may be possible to treat a color transparency as a new original and make 
separations from it in the normal direct manner. 

The processes discussed up to this point are subtractive. Dufaycolor and 
the other screen plates are additive. But, the same end-result is true. It 
must be recalled that in the additive processes the different color densities lie 
in juxtaposition, and occupy a space small enough to register three elements 
as a single unit. Now consider white light falling upon one such elementary 
unit containing three elements, each in a different primary color. The red 
element will, absorb the green and blue rays completely and will transmit the 
red rays to the extent allowed by the silver deposit lying below. Thus only 
red light, and this to a predetermined intensity, is passed. The green and blue 
dots act in a corresponding manner. Through the entire set there will there- 
fore be transmitted intensities which are the same as those reflected from the 
corresponding point in the original subject. Suppose, now, that the illuminat- 
ing light is filtered so that it is a simple primary color, say blue. This will be 
transmitted only by the blue elements and will be modulated by the silver 
deposits lying beneath each blue element. Thus only blue intensities or trans- 
missions will be photographed if the light is made to fall upon a photographic 
emulsion. Once again the colored transparency acts as if it were the original, 
and can be copied in the direct manner discussed above. In this discussion, 
the assumption was made that the colors were pure, and offered no inter- 
ferences. This is only true in theory. The procedures to be used to overcome 
defects arising from the fact that the colors are not ideal, will be discussed later. 

The Finlay plate offers a slightly different problem. Here, it must be 
remembered, the image is separate from the screen. Of course it is possible 
to make a positive from the negative, and register this with a viewing screen. 
In that event, a condition is had identical with those discussed above, and the 
separations can be made in the normal manner. Since it is a positive trans- 



SEPARATION NEGATIVES 203 

parency that is being copied, the separations obtained will be negative in form. 
If the negative is registered with the original taking screen, the transparency- 
is a colored negative that is complementary in color as well as in tone. In 
this case the separations that are made by the use of filtered light, would be 
positives. This system would be recommended where separation positives 
are desired, as in pinatype and in certain photomechanical processes, where 
printing is done from positives. 

Should one examine a Finlay negative closely, it would be seen to consist 
of a checkerboard arrangement of squares, each unit of which has a different 
density. Each square represents the density of a single primary color, and the 
four squares immediately adjacent represent the densities of the other two 
primaries. The Finlay company has prepared two black-and-white screens. 
These are used to block out the squares that correspond to a given primary. 
One screen is used to prepare the magenta and cyan printers, the other for the 
yellow. This is done by registering the negative with the block-out screens, 
so that all the squares with the exception of those relating to one primary, 
become blocked out. When light goes through the combination, it will be 
stopped completely by the block-out screen from going through two of the 
elements, and will be modulated by the silver image of the negative in those 
spots corresponding to the third color-element. 

To make the cyan printer, use the block-out screen marked "Red and Blue." 
It is rather a pity that the Finlay company should fall into the common error 
of calling the cyan, blue, and the magenta, red. The screen is registered 
with the negative so that the outer band of the registration edge at both ends 
appears opaque, and the circles and inner band appear transparent. It is 
then copied by projection upon a process or upon a Hammer Slow or East- 
man 33 plate. To make the magenta printer, register the same screen so that 
the circles appear opaque and the outer and inner bands transparent. The 
yellow printer utilizes the other block-out screen, and this is registered so that 
the inner band appears opaque, while the circles and the outer band appear 
transparent. The block printer is made by binding the negative with a plain 
piece of glass whose thickness is the same as that of the block-out screens. 
It will require, approximately one-sixth of the exposure given the others, this 
being the density of the screens. 

The positives could be developed in any solution capable of giving good 
contrasts. The following formula is especially recommended: 



Sodium sulphite 


37J parts 


Glycin 


15 parts 


Sodium carbonate 


75 parts 


Water to 


1000 parts 



For use, dilute with an equal quantity of water. , Develop five or six minutes 
at 65 F. 
From the description of the working details of the block-out screens, it does 



20 4 HISTORY OF COLOR PHOTOGRAPHY 

not seem to be an extremely difficult matter to prepare them one's self. Con- 
sider the screen to be used to make the magenta printer. It is desired to 
opaque out the red and the blue elements, and to pass completely the green 
elements. This may be done by making a contact print of the taking screen 
upon a panchromatic emulsion, through a red and a blue filter. The red filter 
would opaque out the areas immediately behind the red elements, and the blue 
filter would accomplish the same result behind the blue elements. Develop- 
ment should be carried out to gamma infinity, so that a density of at least 2.5, 
and preferably one of 3.5, will be obtained. The spaces behind the green dots 
should remain completely transparent. Therefore a process pan emulsion 
should be used. The screen and the negative should face each other, hence 
the right and left positions in the screen must be opposite to those in the 
negative. The procedure outlined in the preceding paragraph would not give 
this reversal of direction. 

One method of accomplishing it would be to make master negatives of the 
taking screen by contact, through the three filters. To make the block-out 
screen for the magenta printer, the master negative is made through a green 
filter. If a positive is made from this, also by contact, the areas corresponding 
to the red and blue elements in the taking screen will be opaque; only the green 
will be transparent. The other two screens are made in an analogous manner. 
G. B. Harrison (Eng. P. 434434) disclosed a system of block-out screens, 
but these had been previously described by C. L. A. Brasseur and S. P. Sam- 
polo (Eng. P. 8390/96; U.S.P. 571314), C. L. A. Brasseur (Eng. P. 15185/05, 
710/07; U.S.P. 897815), and G. S. Whitfield (Eng. P. 167793). 

Two types of duplication are possible from screen plates. Reproductions 
can be made upon a similar screen material, or separations can be made by 
means of filters, and the separations then used for the making of duplicate 
prints by any of the positive procedures. A great deal of study has been made 
with regard to both of these types of duplication. Because of the impure 
transmissions and absorptions of the screen colors, it was early determined 
that only sharp-cutting filters could be used (C. L. A. Brasseur; Eng. P. 493 2/07 ; 
Ger. P. 214323 and 219821; Fr. P. 367834; U.S.P. 1163207). When printing 
one screen upon another, an interference pattern known as moire is formed. 
The intensity and visibility of this could be varied somewhat by angling the 
two screens relative to each other, a fact disclosed by A. Lehner (U.S.P. 
1113359; Ger. P. 221916). This phenomenon is well known in the photo- 
mechanical industry, where the halftone negatives are screened at different 
angles to overcome moire. 

The general problem of reproduction from screen plates, was studied by 
C. E. K. Mees {Brit. J. Phot., Vol. 54 (1907), Col. Supp., p. 49). He pointed 
out that when one screen plate is printed upon another, color dilution or degra- 
dation takes place. These findings were corroborated by C. Welborne Piper 
{Brit. J. Phot., Vol. 54 (1907), Col. Supp. p. 81; and by E. Stenger and 
F # Leiber {Brit. 7. Phot., Vol. 55 (1908), Col. Supp., p. 30, 69). More recently 



SEPARATION NEGATIVES 



205 



a detailed study of the problem was made by G. B. Harrison and R. G. Horner 
{Photo. /., Vol. 79 (1939), p. 320. These men concerned themselves with the 
problem of making contact reproductions from Dufaycolor film. 

We will first take up the Mees paper. Color dilution results from the fact 
that it is not possible to register exactly the screen elements in the master 



Tied Green B/tue 



Red 


I J 2 ! 3 


Given 


WMMS§3, 


B/ue 


7 ! 8 1 9 

1 1 



/V EG AT I VE 
FIG. 45 




FIG. 46 



and copy. Consider a line screen (Fig. 45) with red, green, and blue lines that 
is exposed to green light, and developed to form a negative. Therefore the 
green line will be opaque. Now print this negative upon another and similar 
line screen (Fig. 46), but with the lines at right angles to the other. This will 
divide the area into nine squares. In the negative, squares 4, 5, and 6 are 
opaque, 1, 2, 3, 7, 8, and 9 are transparent. Upon superposition of the positive 
we will have the following condition. Square No. 1, transparent in the nega- 

POSITIVE 



Red 


Groan 


Blue. 


m 


2 


3 


4 


5 


6 


7 


8 





FIG. 47 

tive, will transmit red light to the positive, and since it is over a red area an 
opaque deposit will be formed. Square No. 2 will project a red element over 
a green, hence no light will be transmitted to the emulsion below. In the 
positive (Fig. 47), square 2 will be transparent. Similarly squares 3, 7, arid 8 
will be transparent. Square 9 in the negative is a blue, and it is transparent. 
In the positive, it is also a blue, hence blue light will be transmitted to the 
emulsion below, so that this area will be opaque. Squares 4, 5, and 6 are 
opaque in the negative, hence these will be transparent in the positive. Since 
the positive is at right angles to the negative, the green line in it will corre- 



206 HISTORY OF COLOR PHOTOGRAPHY 

spond to squares 2, 5, and 8. These should be, and are, transparent in the 
positive. The areas behind the blue and red lines must be completely opaque, 
since these are transparent in the negative. The blue lines correspond to squares 
3, 6, and 9, and of these only 9 is opaque, the other two being transparent 
because in printing differently colored elements fell upon each. A similar con- 
dition holds with squares 1, 4, and 7 where only square 1 is opaque. There- 
fore instead of three transparent sections, the positive contains seven, of which 
three are green, and two each are red and blue. The end result is a green 
diluted with two parts of white. 

We proceed now to the Harrison and Horner paper. In making copies of a 
Duf ay color film, it cannot be regarded simply as a new original. For one reason, 
the spectral composition of any color in the transparency is no longer the same 
as the original, but a visual match with it. The spectral transmissions of the 
elements, intended as analysis filters, will not act in this manner in additively 
produced reproductions. The result would be a considerable dilution of the 
colors. This would be caused by the fact that the transmissions of the colors 
in the screen have considerable overlaps. Consider a pure green color in the 
original. In the colored negative transparency this would correspond to opa- 
que densities behind the green element, and transparent areas behind the other 
two. When reproduced in the positive, it will be required to form opaque 
areas behind the red and blue elements, and complete transparency behind 
the green. Let us consider the red element. The projection of this part of 
the negative will fall upon not only red, but also upon green and blue elements 
in the copy, since it is not possible to register master and copy with respect to 
the screen elements. If the blue and green did not transmit any red, no image 
formation would take place at these points. But the Dufaycolor blues and 
greens have appreciable red transmission, and vice versa. Hence silver de- 
posits will be formed behind the green and blue squares. In a similar manner 
the projection of the blue element in the negative will cause densities to form 
in the red and green areas. Hence there will be formed an appreciable silver 
deposit behind the green elements in the positive, which will reduce the bril- 
liance and intensity of the green image. This is an effect that is caused solely 
by the presence of overlaps in the spectral transmissions of the screen elements. 
This effect must be carefully distinguished from the one described by Mees, 
and discussed above. 

To make a copy from a Dufaycolor film, no further analysis is required, as 
this has already been achieved during the first exposure. The color record 
consists of a set of three related densities which determine the proportions in 
which the three reseau colors should be mixed, when viewed. In order to 
preserve these proportions in the copy, it is necessary to insure that the light 
passing through one color-element of the first picture can affect the copying 
emulsion only behind elements of like color. 

A second difficulty arises from the regular geometrical pattern of the screen. 
When two elements of like color fall on each other, opaque areas result in the 



SEPARATION NEGATIVES 



207 



positive. When they are different, transparent areas will form. Because of 
the regularity of the patterns, these interferences will form a regular mosaic 
that is distinctly visible and annoying. It is known as moire. 

Both of these disadvantages could be overcome if it were possible to register 
the screen elements of master and copy, but this is not possible. During 
processing, film shrinks slightly, so that the dimensions of the screen elements 
are no longer the same in master and copy. Besides being visible and annoying, 
the effect of moire would be to dilute the colors, as pointed out previously 
by Mees (cf. above). 

If the two screens are separated by the thickness of the celluloid base, and 
if the negative is illuminated by a perfectly diffuse light whose dimensions are 
large with respect to its distance from the negative, the negative reseau will 
become diffuse. The image also will be diffuse to a like degree, and this 
would make the results so poor in definition that they would be unusable. 
The problem is not so much to remove the moire, as it is to reduce it to a 
point where it will no longer be noticeable, and this with a minimum loss of 
image definition. 

In printing by contact, it is essential to have the negative and positive 
separated from each other by the thickness of the base, in order to allow the 
light that passed through the master to go through the reseau of the copy, 
before acting upon the positive emulsion. This separation between reseau 
elements can be utilized to effect complete removal of the master-screen 
structure. Consider first of all a pinhole camera (Fig. 48). The image B of 
the object A is a perfect reproduction, in so far as relative dimensions are 
concerned. The size of the reproduction is determined solely by the distance 




FIG. 48 



of the object from the pinhole, and the size of the object. The Duf ay color screen 
may be regarded as being composed of a number of juxtaposed squares each of 
which contains one red, one green, and one blue element. Each of these is of a 
dimension sufficiently small to act as a pinhole camera. Now consider the set- 
up in Fig. 49. The distance d is the length of one elementary cell in the reseau, 
and since there are 500 to the inch, this will have a dimension of 0.002 inch or 0.05 
mm. The opening in the negative reseau has the dimensions of one of the color 
elements which make up the elementary cell, and which acts as the pinhole 
for the pinhole camera. D is the dimension of a uniformly lighted opal glass. 
The pinhole will image this upon the positive reseau. It is desired to make 
this image the exact size of the cell unit, which is equal to 0.05 mm square. 



208 



HISTORY OF COLOR PHOTOGRAPHY 



r resecut 



e/eme/its * 



negative 
resea,u 




FIG. 49 



Since the size (d) of the desired image, and the distance (t) between the two 
reseau patterns are constant, the ratio between D and T becomes fixed. The 
following relationship holds between the values d, t, D, and T. 



D 
T 



A. 



\xd 

T 



where fx is the index of refraction for the celluloid base. Assuming d = 0.05 mm, 
t — 0.005 mcn or °- I2 5 mm > M = I -4°> an d T = 10 cm = 100 mm = 4 inches, 

^ 1.4 X 0.05 X 100 , ... 

D = - — — — = 56 mm = 2x inches 

0.125 * 

Hence the light source would have to be 2 \ inches square, situated a distance 
of 4 inches away from the master and copy, which are in contact with each 
other. Under these conditions each color element in the master will be imaged 
over the entire unit of three color elements in the copy. The adjacent color 
element will likewise be imaged over a full set, so that this scheme provides 
the exact amount of diffusion required to convert the screen into a continuous 
tone. Moire is therefore completely eliminated. 

In order to destroy moire, it has been necessary to spread each color element 
in the master so that it covers an area containing three elements in the copy. 
Thus red light falls not only upon a red element in the copy, but also upon 
blue and green elements. Since it is required that a red element print only 
behind a red element, it is essential to examine the effect of red light trans- 
mission through the blue and green elements and vice versa. 

The spectral transmissions of the screen elements are far from being mutually 
exclusive, and the colors are not completely saturated. The reasons for this were 
discussed in The Photographic Journal, 1937, p. 706. Because of the overlaps in 
transmissions, when a red color element falls upon a green element, an appreci- 
able quantity of light is transmitted to the emulsion below, thus giving rise to 
silver deposits at these points. In making copies, it is this overlapping light 
that must be kept to an absolute minimum, to reduce color degradation and 
falsification. 

Let P x be the sensitivity of the emulsion to light of wavelength X, lf x 



SEPARATION NEGATIVES 209 

the intensity of this wavelength in the printing light, and P x , Gx, and P x 
the transmissions of the red, green, and blue reseau elements to light of this 
wavelength. Then P x 2 , P x 2 , and G x 2 will be the transmissions at the points 
where the color elements are in registry; that is, where a red falls upon a 
red, a green upon a green, and a blue upon a blue. And P X G X , P X P X , and 
Px^x will be the transmissions of the overlapping portions, where a red 
falls upon a green, etc. We have then six cases, as follows: 

1. Red — Red 4. Red — Green 

2. Green — Green 5. Red — Blue 

3. Blue — Blue 6. Green — Blue 

The photographic effect of light of wavelength X upon the emulsion below, will 
be given by the product 

Mx • Px • X x • F x 

where X x and F x are the terms P x , Gx, or P x . Where the red elements coin- 
cide, the expression becomes 

Mx - P x • P x 2 

Where the red overlaps the blue, the expression becomes 

Mx • Px • Px • Px, etc. 

To find the effect for the whole spectrum, these expressions must be integrated 
over the entire visible range, thus: 

X700 
m Mx - Px • Px 2 • d\ = 100 (arbitrary) 

2. Green — Green = | Mx • Px * Gx 2 • dA = 100 (arbitrary) 

J 400 

3. Blue — Blue = l^x ■ Px • Px 2 • dA = 100 (arbitrary) 

J»700 
MxPx'Rx'Gx'd\ = 18.3 
400 

/•700 

5. Red — Blue = | M x • P x • P x • P x • d\ = 15.7 

J 400 

/•700 
MxPx-Gx-Bx-d\ = 23.8 
400 

The values for the cases where the colors coincide have been taken arbitrarily 
as 100, and equal to each other. This is not strictly true, but serves our 
purpose. The other values are those obtained when Mazda light is used for 
printing. The overlap of the red into the green is therefore 18 per cent. The 
blue elements have appreciable red transmission from 630 m/x on. The over- 
lap in this case is almost 16 per cent. The worst offender is the blue overlap 
in the green, 24 per cent. This amount of desaturation cannot be tolerated, 
hence precautions must be taken to reduce them to a minimum. 



210 HISTORY OF COLOR PHOTOGRAPHY 

A glance at the term 

. Mx-P^Rx'Bx 

which represents the amount of action due to overlapping colors indicates 
that only M x and P x can be varied, since R^, B\ and G x are fixed by the 
spectral characteristics of the reseau colors. It remains, therefore only 
possible to make the product M x • P\ a minimum in the range of the over- 
laps. Since P x relates to the photographic emulsion it becomes much simpler 
to vary M x , the spectral characteristic of the printing light. This can be done 
by means of filters. The overlap in the deep red, due to a dichroic transmission 
of the green and blue beyond 650 m^u, can be overcome by limiting the sensi- 
tivity of the emulsion to 650. This is the sensitivity boundary of the normal 
panchromatic emulsion. It is only the specially sensitized and the infrared 
emulsions that have sensitivities beyond this. 

The best method of changing M\ in the desired ranges is to print through 
sharp-cutting filters whose transmissions are given by lf x R , M X G , and M X B . 
The Dufaycolor S and P filters are such. Using the sharp-cutting filters, 
the following conditions exist: 



Color Combination 


Red Filter 


Green Filter 


Blue Filter 


Red — Red 


100 


9.6 


54 


Green — Green 


2.5 


100 


2.5 


Blue — Blue 


2.0 


4-5 


100 


Red — Green 


6.7 


4.1 


2.3 


Red — Blue 


4.8 


34 


4-5 


Green — Blue 


2.5 


5.8 


7-3 



This represents a considerable increase in efficiency over the use of unfiltered 
light. 

The regions in which the colors of the reseau elements are completely free 
from overlaps, are very narrow. Even with the use of monochromatic light, 
it is not quite possible to completely eliminate them. But a very good ap- 
proximation can be obtained by the use of the red line of cadmium (X = 6438A) 
for the red, the mercury line (X = 5461) for the green, and the two mercury 
lines (X = 4047, 4358) for the blue. The yellow mercury doublet (X = 5780) 
can be removed by use of a didymium filter. This absorbs strongly in the 
yellow, but transmits freely in the blue and the green. Since the use of the 
cadmium line is but slightly better than the use of the sharp-cutting red filter, 
and since the filter is so much more practical, it is inadvisable to resort to the 
very low luminosity occasioned by the use of a cadmium lamp. 

The article in The Photographic Journal covers printing from Dufaycolor film, 
rather completely. It is surprising that no mention has been made of the 
previous work in this field. The problem of moire was first discussed in a 
rather detailed manner by Mees, Piper, and Stenger and Leiber, as was pointed 
out above. Von Hiibl also studied screen-plate printing. His work is reviewed 
rather fully in Wall's "History of Three-Color Photography" (1925), page 542. 



SEPARATION NEGATIVES 



211 



The original articles appeared in the Wiener Mitteilungen (19 10) and were ab- 
stracted in the British Journal of Photography, Volume 57 (1910) Colour 
Supplement, page 59. 

When an Autochrome plate is printed so as to obtain a print in mono- 
chrome, the results are weak, since the light affecting the color-blind emulsion 
must come only through the blue elements. Hence, even in the deepest shad- 
ows, the blacks are diluted with two portions of white. The same difficulty is 
met when it is printed on another screen plate (since the chances of a blue 
dot falling on another blue dot are one in three). It is possible to obtain sharp 
images from Autochromes even if the emulsion side of the copy is placed 
against the glass side of the Autochrome, this occasioning a separation equal 
to 2 mm between them. The two are loaded into a camera (the lens of which 
is set at infinity so as to obtain parallel rays), and the camera then pointed 
to the sky, which must be CQvered with white clouds. This condition can also 
be duplicated in the laboratory, merely by exposing the combination in the 
camera to a large white area illuminated perfectly evenly. The image of this 
area must be sufficiently large to cover the entire surface of the Autochrome. 
Another way that this can be accomplished is to use a point source of light 
situated at the focal point of a set of condenser lenses whose diameters are 
larger than the diagonal of the Autochrome. The Autochrome plus copy 
material, is placed adjacent to the other side of the condensers. . 

Although procedures such as the above overcome the problem of definition 
loss due to lack of contact, the problem of color rendition still remains. Con- 
sider the case where a certain red color is composed of 300 transparent and 
600 opaque grains. In copying this upon another screen plate, only 100 of 
the red grains will fall upon other red grains. The remaining grains will fall 
upon blue and green elements. No light will pass through these combinations. 
Therefore, in the copy, the reproduction will consist of 100 transparent 
grains and 800 opaque ones. If advantage be taken of the geometry of the 
situation, this can be overcome, although with a slight loss of definition. 
Consider first of all the case of a point source of light, as in Fig. 50. The 




FIG. 50 



FIG. 51 



212 HISTORY OF COLOR PHOTOGRAPHY 

elements 5 (opaque), and e (transparent) of the master screen will be repro- 
duced sharply in the copy. But if the source of light is not a point, a different 
story is true, as is made evident in Fig. 51. Now the image e h of the transparent 
portion e in the master, will be enlarged, the degree of enlargement being d. 
This is also a measure of the unsharpness that is introduced by this method. 
From the geometry it is seen that conditions can be chosen so that the image 
ei is enlarged to cover a whole unit consisting of a red, green, and blue element 
in the copy screen. Then, if e represents a red dot in the master, it will have 
to fall upon a red dot in the copy, since its image, d, will encompass all three 



jf~e> 




FIG. 52 

colors. Instead of enlarging three times, which will be just sufficient to cover 
the spaces between the dots in the copy, Von Hiibl suggests a four-time en- 
largement. Consider the nature of the image e h in its greatly enlarged form 
(Fig. 52). It is only the central portion that receives light from every point 
on the light source L. The edges, therefore fall off in intensity, hence the 
image ei of the element e is not uniform. But if the degree of enlargement is 
such that the edges overlap each other, their intensities will be increased. 
At a four-time enlargement this is sufficient to give a completely uniform 
density in the copy. Von Hiibl made a mathematical analysis of the problem, 
much in the same manner as was done by Harrison and Horner, twenty-nine 
years later. He obtained the equation: 

r A{V ~ I) 

Ju = • e 

a 

where L is the size of the light source, 

A is the distance the light source is from the material, 

a is the separation between master and copy, 

v is the degree of enlargement desired, and 

e is the dimension of the dot in the master. 

He also determined the degree of unsharpness, d. This is related to the other 
dimensions by the equation 

d = —a 
A 

It has often been proposed to diffuse the image by racking the lens slightly 
out of focus. This will cause a slight diffusion which should be sufficient to 



SEPARATION NEGATIVES 213 

cause the screen elements to fuse. J. Tritton made use of the poor resolving 
power of the lens to effect a fusion of the elements. It is a well-known fact 
that when light passes a sharp edge, it becomes slightly diffracted. The 
diaphragm of a lens is also a sharp line, and the light passing through the lens 
will be affected in this manner. If the circumference of the lens opening is 
made small, the diffusion caused by this phenomenon will be sufficient to fill 
in between the images of the elements. When the copy is made one-to-one 
{Brit. J. Phot., Vol. 84 (1937), p. 513) the aperture must be closed to at least 
/: 4 5or/:64. 

The interposition of substances into the paths of the projected image to 
give a smoothing effect, has been a popular remedy. G. Heymer (U.S.P. 
1879236) proposed to place a lenticular film base in the path. The lens ele- 
ments of the film are such as to cause interference patterns which destroy 
the moire. W. Chapman (U.S.P. 1955715) placed a doubling image prism 
in the path of the rays. The two images are spread sufficiently to fill in the 
space between screen elements. This is extended to a quadruple image in a 
subsequent disclosure (U.S.P. 2031032; Eng. P. 389345). The same inventor 
also proposed to interpose in the path of the rays an element which will cause 
diffusion to take place, and in that way cause a fusion of the screen elements 
(U.S.P. 2049556). T. T. Baker accomplished the same result by interposing 
a sheet of glass with linear or circular depressions, 0.1 mm in depth (Eng. P. 
337041). The same result could be accomplished according to T. T. Baker 
(U.S.P. 1903971; Eng. P. 366958), if a lens be used at an aperture of at 
least/: 2. The lens is focused sharply upon the silver image of the master film. 
The depth of focus at this opening is so slight that the screen in the master 
will be sufficiently diffused to be negligible. This procedure is especially 
useful in copying a screen plate upon another screen plate. 

F. A. Lindemann proposed to use two light sources at a slight angle to each 
other, and copy the screen plate by contact, but the emulsions in master and 
copy are placed a definite distance apart. The angularity of the two light 
beams is such that they cast adjacent images, thus filling in the spaces between 
the screen elements (Eng. P. 374891). Or a single light source could be used 
in front of a bi-prism, to yield the two beams. The method was originally 
proposed to overcome the moire patterns formed when printing from a screen 
master to a screen copy film, but the application to separations is immediate. 
A somewhat similar idea is disclosed by S. D. Threadgold (Eng. P. 446679). 
He used a diffuse light source, and this gave an enlarged image, just sufficient 
to cover three elements. Thus the space between the elements becomes filled 
in. This is no different from the original von Hiibl suggestion, or the Harrison 
and Horner technique. Threadgold preceded the last mentioned disclosure 
by several years. 



CHAPTER 16 
THE LENTICULAR PROCESS 



I 



N this chapter we turn our attention to one of the simplest and most elegant 
of all the proposals for the making of color reproductions. This is the lenticu- 
lar process, which requires no special camera or special processing technique, 
except that a special type of film has molded on its back a series of lenticular 
elements which divide the surface into a honeycomb structure. For a long 
time after the idea was first proposed, nothing was done with it. But later it 
became extremely popular, and for a time it was thought that this scheme 
solved all problems of motion pictures in color. Some idea regarding the posi- 
tion which lenticular film held in the industry may be obtained from the 
fact that fully one-third of the patents dealing with color reproduction in 
the decade 1925-1935, dealt in one way or another, with this process. This 
interest waned considerably with the advent of the monopacks. 

The principles which govern the use of lenticular film for color reproduction 
purposes, are purely optical in nature. Therefore, it may be of value to re- 
view at this point the fundamentals of lens and camera optics. Consider the 
following situation. An object is being photographed at a distance of ten 
feet, with a 4 by 5 camera, equipped with an 8-inch lens, the exposure being 
made at an opening of /: 8. The lens will have an effective aperture that 
measures one inch in diameter. From every point on the object there will 
radiate a cone of light, and the camera lens will intercept that portion of it 
which will have a cross section one inch in diameter (Fig. 53). Let us center 
our attention upon the cone of light which enters the lens. It has its apex 
on the object being photographed, 120 inches away from the camera. The 
base of the cone is a circle one inch in diameter. Therefore the angularity 
of the cone is very small. To all intents and purposes, it can be considered 
that the rays entering the lens are parallel to each other, and that their point 
of convergence is an infinite distance away. Stated a little more scientifically, 
ten feet represents the infinity distance in this case. From every point on the 
object, then, there arises a beam of parallel rays, each beam making a different 
angle with the axis of the lens. That is the only difference between all the 
rays which enter the lens, and it is this difference which determines exactly 
where the image of any point will appear in the negative. But this is getting 
a little ahead of the story. 

Lenses are possible because light travels at a different rate of speed through 
dense transparent substances, such as glass, than it does through air. The 

214 



TEE LENTICULAR PROCESS 



215 



ratio between the velocities of light in vacuum and in any other medium, is 
called the index of refraction of the material. Now consider the case of 



■/ens 




lutck 

camera 



object 



FIG. 53 



a wave front AB, traveling in air, then entering a glass plate whose outer 
surface is depicted by the line CD in the diagram (Fig. 54). Let v be the 




G 



FIG. 54 



216 HISTORY OF COLOR PHOTOGRAPHY 

velocity which the wave front has in glass, and C its velocity in air. The 
wave front is traveling in the direction EE\ which is normal or perpendicular 
to AB. Consider the wave front from the moment the point A on it reaches 
the glass. As the beam moves forward, the point A moves into the glass 
where it travels with a velocity v. In a time /, the distance it will have traveled 
in the glass will be vt = AA'. During that same time, the wave front will 
have traveled in air a distance Ct = EE' } and the point E on the wave front 
will have reached the glass surface. Inside the glass, the wave front will 
take the form E'A f G y and the direction of its motion will be along a line 
perpendicular to the line E'A'G* 

The direction of the original beam has therefore been changed when it 
passed from air to glass, as indicated by the line EE'F'. This break in the 
direction of a beam of light when it passes from one optical medium into the 
next, is called refraction, and the amount of change in this direction is measured 
by the index of refraction. It is quite easy to establish from geometrical con- 
siderations that the index of refraction is numerically equal to the ratio of 
the velocities in the two media. Glass, for instance, has an index of refraction 
ranging approximately between 1.50 and 1.70. The film base has a value 
of 1.40. Pure gelatin also has a value of 1.40 or thereabouts. 

When a beam of light goes from air to a denser medium, its direction is 
always bent toward the normal to the surface of the denser medium. When 
a beam goes from a dense medium to air, it is always bent away from the 
normal to the interface between the two media. There is a cardinal rule in 
optics, the complete reversibility of action, and it is the one guiding principle 
in the entire study of the lenticular processes. Therefore the wave front 
E'A'G, upon emerging from the dense medium into air, will undergo the 
exact opposite direction transformation, provided the interface D'C is optically 
equivalent to the interface CD. This is true if the face D'C is parallel to the 
face CD. We can state the very important principle that is utilized in all 
beam-splitting devices, that when a beam of light passes through a sheet of 
glass with parallel faces, the direction of the beam is not changed, but its 
position is somewhat displaced. Thus in our diagram, were no glass plate 
included in the path, the direction of the wave front would be along the 
line EE'FR'. But upon placing the glass plate in the path, the beam pro- 
ceeds along the broken line EE'FR. The line F'R is parallel to FR' so that 
the direction is the same as previously, but the position has been displaced a 
distance FF'. 

When a beam of light (this can be defined as the normal to the wave front, 
hence could be identified with the direction of the wave front) goes from air 
into a glass prism and then emerges from the prism, an entirely different story 
is true (Fig. 55). Consider the glass prism ABC, and the beam of light RD. 
Upon striking and entering the prism, the beam becomes refracted, and 
travels along the line DE until it reaches the new glass-air interface. Here 
the side A C is no longer the optical equivalent of the side AB y since it is not 



THE LENTICULAR PROCESS 



217 



parallel to it. Hence upon emerging, the beam will be deflected away from 
the normal to the surface AC, and the new direction will be along ES. This 
is no longer parallel to, nor does it have any relationship to, the direction of 




FIG. 55 

the line RD, but is dependent mainly upon the angle of the prism, BA C. 
It is seen, therefore, that by properly designing a prism, it is possible to direct 
the beam to any desired place or direction. 

Consider, now, a lens LL' (Fig. 56). This is a transparent substance whose 
outer faces have been polished down so that each point on the surfaces offers 




FIG. 56 

a different angle to a beam of parallel rays. Such a beam is illustrated by the 
rays R, S, T, U, V, W, all of which are parallel to each other. If the surfaces 
of the lens are spherical, each of the rays, R, 5, T, etc., will strike the front 
surface at a different angle. The ray RA will be refracted to the point A' 
by its passage through the lens. The angle which the refracted ray A A' 
makes with the other lens surface is different from the angle it makes with 



2l8 



HISTORY OF COLOR PHOTOGRAPHY 



the first, hence upon going from the- lens to air it will not proceed along the 
line parallel to RA, but along a different direction A'R'. A similar condition 
will be true for the ray WF. After passage through the lens, it will proceed 
along a direction F'W, which is not parallel to A'R\ since the surface 
LABCDEFL' is not parallel to LA'B'CD'E'F'L'. The curvatures of the 
surfaces can be so chosen that the rays, after emerging from the lens, will 
intersect at some point, M. It has been found that if these surfaces are spheri- 
cal in shape, and if the centers of these spheres lie on the same straight line, 
then a beam of parallel rays entering the lens will emerge in the form of a 
circular cone whose apex is at the point M. The line which joins the centers 
of the two spherical surfaces of the lens, is known as the optical axis (Fig. 57). 
From the geometry, this line must be normal to both surfaces at the points 
of intersection with them. A ray of light traveling along this path, will enter 



M 




FIG. 57 



and leave the glass medium normally, hence will not suffer any change in di- 
rection. If a beam of rays, all of which are parallel to the optical axis, enters 
the lens, it will converge at a point M which lies on the axis. This point is 
called the rear focal point of the lens. The plane through this point, per- 
pendicular to the optical axis, is called the rear focal plane, or the image 
plane of the lens. This plane plays a very important role in photographic 
optics. The negative material is placed in this position. 

If two beams of light, each consisting of rays that are parallel to each 
other, enter the lens from different directions, each beam will be brought to 
a focus at a different point. This is illustrated in the diagram in Fig. 58. 
The system of parallel rays jBi, B2, B3, enters the lens in the form of a beam, 
and leaves it in the form of a cone B\ B 7 ' B', with the apex at B', the point 
where all the rays B\ to B-i come to a focus. In a similar manner the system 
of rays A \Ai enters the lens as a beam but leaves it as a cone, coming to a focus 
at the point A'. The locus of the points A* and B f is determined solely by 
the angle which the rays make with the optical axis, RR f . The ensemble of 
all the points A\B\ . . ♦ ', which are the apices of the cones that result when 
beams of light enter the lens at all possible angles, forms a plane which coin- 
cides with the rear focal plane or image plane of the lens. 

The position of a point in this plane is determined by means of two entirely 
independent numbers. The direction which any line in space makes with a 



THE LENTICULAR PROCESS 



219 



given standard direction is determined by three numbers. But these are 
connected by a relationship which states that the sum of the squares of these 
numbers (called direction cosines) must equal one. Hence only two of the 
numbers are independent. It is not a difficult problem for a mathematician 
to develop the connection between the two sets of numbers. 




FIG. 58 

A beam of light with parallel rays A, B,C, . . . H, entering a lens (Fig. 59) 
is brought to a focus at a point 0, whose position in the image plane is de- 
termined by the angle which the beam makes with the optical axis RR'. The 
beam, after passage through the lens, will take the form of a cone of rays. 
In this cone, there will be one, and only one, ray whose direction is the same 
as that of the beam, and which will not suffer a change in direction by passage 




FIG. 59 

through the lens. This is the beam D'. It intersects the optical axis at the 
point N, which is called the nodal point of emergence of the lens. The ray 
D f in the emergent cone corresponds to the ray D in the beam. This ray 
intersects the optical axis at the point M, which is called the nodal point of 
admission. Every beam of light, containing rays that are parallel to each 
other, will have one ray which proceeds from the source directly to the point O. 
The beam upon emergence from the lens, will appear as if it had emerged 
from the point N, and will have a direction parallel to that of the beam. 
The two nodal points M and N have a very important geometric property. 



220 HISTORY OF COLOR PHOTOGRAPHY 

Draw a line from M in any direction, so that it intersects the entry surface 
of the lens RS (Fig. 60). This will be at the point B. At this point draw 
the tangent to the lens surface, BA. Now draw a line parallel to MB from 
the point N, to the emergent side of the lens. This will intersect the lens 
surface at C. Draw the tangent to the surface at C. This will be the line CD. 
The line CD will be parallel to the line AB. Now consider the ray LBM, 
whose direction is identical with the direction BM. This ray upon entrance 
into the lens will be refracted toward the normal to the surface at the point B. 
The refracted ray will pass through the lens to the point C, at which point 




it goes from glass to air, hence will be refracted away from the normal to the 
surface at the point C As far as the ray LB is concerned (appearing to go 
through the nodal point of admission M) the lens is merely a glass plate 
with parallel sides AB and CD. Hence the ray will emerge in the same di- 
rection that it entered, but it will be displaced somewhat and appear as if it 
originated from the nodal point of emergence. 

The existence of nodal points makes geometrical optics quite simple. Every 
lens can be classified exactly if the positions of the two nodal points and the 
rear focal plane be given. It has been noted above that from every point on an 
object being photographed, there radiates a cone of rays. That section of 
the cone which enters the lens aperture can be considered to consist of a 
group of parallel rays, hence the lens will bring these to a point focus some- 
where in the focal plane of the lens. Among all the rays coming from a single 
point on the object, there will be one ray R, that will pass through the point M 
in the lens system, the nodal point of admission (Fig. 61). This ray will then 
translate itself from M to N y the nodal point of emergence, and continue 
from there in a parallel direction NR', until it intersects the focal plane F 
at the point P. This is the image point corresponding to the point on the 
object from which the beam characterized by the ray R, originated. In this 



THE LENTICULAR PROCESS 



221 



manner, every point on the object gives rise to a single point in the image or 
rear focal plane. It must be understood that the discussion up to this point 
relates only to perfect and ideal optical systems, which serve as the goal 
toward which lens designers and opticians strive. Lenses are complicated 




7WK5 fro* 
poutt o/t ot>j 



FIG. 61 



affairs, the complications arising from the attempts to make them appear as 
close as possible to the ideal. It is only within the very limited and precise 
limits in which the commercial lens approaches the ideal, that they can be 
used; and one should never use them outside these limits, unless one is willing 
to accept poor definition and resolving power, or other inferior quality. 




FIG. 62 



Consider a light source at a point P (Fig. 62) which lies in the rear focal plane 
of a lens. This will radiate light rays in all directions and, in particular, will 
fill a lens aperture with a cone of light composed of the rays A y B, C, D y E } 
and F. One of the most important properties of the lens is that the cone of 
light; originating at a point in the focal plane of a lens will emerge from that 
lens in the form of a beam containing parallel rays, which come to a point 



222 HISTORY OF COLOR PHOTOGRAPHY 

focus at infinity. This is the exact reverse of what happens when a beam of 
parallel rays enters the lens. There is therefore a complete reversibility in 
optical phenomena. This is the principle that makes lenticular processes 
possible. 

Now let us consider the application of these principles to lenticular film. 
It was Gabriel Lippmann, the brilliant scientist who also developed the 
grainless Lippmann emulsion, who first suggested the possibility of a lenticular 
film material (CompL rend., Vol. 166 (1908), p. 446; Brit. J. Phot., Vol. 55 
(1908), p. 192). He suggested that a film base be impressed with a honey- 
comb structure, each element of which would act as a tiny lens. This will 




FIG. 63 

image only that portion of light which is directly in front of it. In particular, 
it should be possible to make these lenses so small that they will see only the 
light that comes from a single beam. Thus in Fig. 63, the beam of parallel 
rays A\, enters the lens L, and emerges in the form of a cone, coming to a 
focus at the point M. It is desirable to make the honeycomb structure on the 
base of such dimensions that each element of the honeycomb will receive 
the light from but a single cone, hence from but a single point on the object. 
The tiny lens at M will then form an image on the emulsion lying directly 
beneath it, of the intensity of the beam that enters the lens at that particular 
angle that will come to a focus at the point M. The entire picture would 
consist of a series of minute dots lying in juxtaposition, each dot being the 
image seen by the tiny lens immediately in front of it. If the dots are beyond 
the resolving power of the eye, they will not be seen as individuals, but as 
blends. 

This description falls short of expounding a color process. To accomplish 
this each of the minute lenses must image three dots, one for each of the 
primary colors. It was R. Berthon, an astronomical optician, who showed 



THE LENTICULAR PROCESS 



223 



how this could be done, just one year after Lippmann made his disclosures 
(Eng. P. 10611/09). Consider a beam of parallel rays A1A2B1B2C1C2, etc., 
entering the aperture of a lens L, in whose nodal plane of admission there is 
placed a special filter (Fig. 64). This consists of three bands lying side by 
side, each dyed the color of a single primary. Together, the three filters com- 




filter In hccUlL^ 

of admission, 



FIG. 64 



pletely fill the lens aperture. The spectral quality of the beam A\, . . . C% 
is uniform, but only part of the beam will go through each portion of the 
aperture. Thus the part A\Az will go through the red, the part BiB 2 will 
go through the green, and the part C1C2 will go through the blue filter area 
of the aperture. After passing through the lens, the beam will be converted 



bUce, cone. 




FIG. 65 



into three cones, each of which comes to the same point focus, M, in the 
image plane. The cones will differ from each other in spectral quality, since 
they have passed through different filters. They will also differ in intensity, 
since the original beam had different intensities for the three primaries present. 
Now consider the three colored cones from the point of view of the minute 
lens situated at the point of focus of the beam, that is, at the point M\. This 
is illustrated in highly exaggerated form in the diagram in Fig. 65. The cone 
of red rays originates from the point on the object and passes through the red 



224 HISTORY OF COLOR PHOTOGRAPHY , 

filter in the aperture of the taking objective, which sees it as a beam of parallel 
rays. The tiny lens, Mi, situated in the rear focal plane of the taking ob- 
jective, will image this cone in the form of a dot that covers one-third of the 
area on the emulsion immediately behind Mi. The density of the dot will 
depend upon the intensity of the red rays present in the original beam. The 
cone of green rays lies directly adjacent to the red, so that the lens Mi will 
image this as a dot lying next to the red density image. The blue will in a 
similar manner be imaged adjacent to the green. The three dots together 
fill entirely the area directly beneath each tiny lens, and so, adjacent to the 
blue dot on the other side, will be the image of the red rays coming from an 
adjacent lenticule, M%. These cones coming to a focus at M2, originate from 
a point of the object that lies adjacent to the first. 

Since the three dots arising from each beam take up a definite area, this 
becomes the unit of definition of lenticular processes. It is possible to make 
film with 20 to 30 such elements to the millimeter, or from 500 to 750 to the 
linear inch. It is seen, therefore, that the effect of a lenticular film, used in 
conjunction with a banded filter in the nodal plane of admission, is identical 
with a screen plate. In the first case the screen is formed optically, whereas 
in the second it is formed mechanically. It is considerably easier and cheaper 
to mold a lenticular film base than to rule or print colored elements. There- 
fore in this limited region, lenticular film has an advantage over screen plates. 
But this is offset by a large number of other disadvantages which served to 
sidetrack the lenticular processes as soon as screen plates and other color 
processes that were not too complicated, came into being. 

Although Lippmann first suggested the possibility of lenticular film, and 
Berthon described the first successful application of the material to color, a 
somewhat similar idea had been current ever since 1895. F. N. Lanchester 
may be credited with the germ of the idea (Eng. P. 16548/95). A ray of light 
passing through a small opening in an opaque screen, and then through a 
prism, will be spread out into a spectrum. If instead of a small opening, the 
light passes through a slit with a definite width, the image will be in the form 
of a rectangular band of color. Now suppose that the slit be divided into 
three parts, and each part covered with a filter passing but one primary, then 
the image will consist of three bands colored red, green, and blue respectively. 
If these be projected upon a panchromatic film, there will be formed three 
silver densities corresponding to the intensities of each primary present in 
the beam that entered the slit at the given angle. In photography by this 
means, we replace the slit with a line screen. 

The germ of the idea disclosed here is that in a screen that is placed in the 
rear focal plane of a lens system, each element acts as a pinhole camera. This 
breaks up the object into minute units, and each unit becomes photographed 
separately by means of a pinhole camera. Mr. Lanchester went on to disperse 
this light by means of a prism, achieving color photography by the micro- 
dispersion method. This has been described in detail in Chapter 3. Here 



THE LENTICULAR PROCESS 225 

we are interested in the one step of placing a screen in the rear focal plane of a 
taking objective, and having the elements of the screen act as individual 
pinhole cameras. 

A further step in this direction was taken by R. E. Liesegang, who was 
apparently the first to suggest the use of banded filters in the aperture of a 
lens system {Phot. Arch., Vol. 37 (1896), p. 250; Brit. J. Phot., Vol. 43 (1896), 
p. 569). In this article, he suggested the use of yellow, red, and blue filters. 
It is rather surprising that so eminent and accomplished a photographic 
scientist as Liesegang should make so elementary an error as to confuse the 
subtractive or secondary colors with the primaries. Since the terms red and 
blue are used to designate two elements of each set, we can recognize which 
set is meant by the third color. Since Liesegang uses yellow, his red and 
blue must have been the red and blue of the printing inks, therefore really 
magenta and cyan. Although this set of colors, magenta, cyan, and yellow, 
does form a balanced unit, it is just exactly one hundred per cent wrong to use 
them for analysis purposes. Each of these transmit two primaries, hence 
there will be no true color analysis. If he actually meant to use primary red 
and blue for two of his colors, then the use of yellow for the third color would 
again have destroyed true color analysis, for instead of obtaining a true 
evaluation of the green intensity, he would have obtained a green density 
mixed with an equal quantity of red. 

But this error in the choice of colors does not mitigate against the sound- 
ness of his engineering and optical procedures. He writes in substance, as 
follows: Every element of a cross-lined screen acts as a pinhole camera, and 
reproduces an image of the aperture of the objective in whose rear focal plane 
it is placed. Thus, when using a square stop, the dots in the halftone produced 
will be square in shape; and when the stop is triangular, the dots produced 
will be triangles. If a diaphragm is used with several holes, each element of 
the screen will reproduce as many dots as there are holes in the diaphragm. 
If these holes be covered with red, yellow, and blue filters, the three dots 
will correspond in density to the intensities of the colors entering the lens 
aperture. In order to convert the negative into its natural colors, it can be 
converted into a transparency and projected through the three-color diaphragm 
and the original lens system. 

The following year, J. A. C. Branfill {Brit. J. Phot., Vol. 44 (1897), p. 142) 
suggested the use of four openings in the diaphragm. Three of the openings 
were covered with the primary filters, and the fourth was left clear. The 
sizes of the openings were adjusted to yield balance by a simultaneous ex- 
posure. Evidently Mr. Branfill had the photomechanical industry in mind, 
since it is here that a fourth (black) printer is demanded. The same idea 
was disclosed by Giesecke (Ger. P. 117598); Szczepanik (Eng. P. 7729/99); 
F. E. Ives, (U.S.P. 648748); and J. de Lassus Saint-Genies (Fr. P. 459566). 

A further advance was made by Edward Russell Clarke (Eng. P. 10690/02). 
He disclosed a method for the making of separations from the integral negative 



226 HISTORY OF COLOR PHOTOGRAPHY 

prepared as above. This was done by preparing block-outs made by exposing 
a plate to a uniform white area, with all but one of the openings closed. From 
such a negative, a positive was made. This had as solids all the area on the 
plate except those portions corresponding to the positions where the image 
of the one open stop would fall. If this were registered with the negative, it 
would block out all densities except the one corresponding to a single opening 
in the diaphragm, hence limit it to the densities of a single primary separation. 
In this manner, by the preparation of one block-out screen for each opening 
of the diaphragm, it is possible to isolate the dots corresponding to each color 
separation. It is also possible to project the negative through an identical 
optical system, and close all but one of the diaphragm openings. The resultant 
image will correspond to the densities of but a single primary. 

From the foregoing it may be seen that the Berthon system was merely the 
optical analogue of the one described by Liesegang, and elaborated by Branfill 
and the other persons just mentioned, and fully described at least ten years 
before either Lippmann or Berthon. These men simply did optically what the 
others have done mechanically. Both of these schemes can be considered 
as the optical equivalents of the screen methods proposed by du Hauron, and 
first put into practice by Prof. Joly. The pinhole camera or lenticular lens 
placed in the rear focal plane of an objective in whose front nodal plane is 
positioned three niters, projects a color screen upon the emulsion. The in- 
dividual cell densities of this screen are determined by the composition of 
the light reflected from the object being photographed. In the screen plate, 
the cross-lined screen consists of elements each of which is colored in a dif- 
ferent primary. In so far as the emulsion receiving the impression is con- 
cerned, the three are identical. It is interesting to note that Liesegang pro- 
posed his scheme frankly as a modification of the Joly screen process. 

Lenticular film achieved a small measure of popularity in the United States. 
About 1928, the Eastman Kodak Company introduced it under the name 
Kodacolor, for amateur 16 mm movies. With the introduction of Kodachrome 
somewhat later by the same company, the lenticular process disappeared. 
It had a life of approximately seven or eight years. Its introduction may be 
ascribed to the terrific pressure that was being applied by the amateur for a 
color process, and it lasted only until a simpler procedure was introduced. 
Although the fidelity of color reproduction was quite high, the process had 
many inherent flaws that made its further use a matter of great inconvenience. 
These will be discussed at a later time. 

Kodacolor was a transparency process, wherein the original film that was 
exposed in the camera was processed to yield a positive. By projection through 
a suitable lens, containing a three-banded filter in its aperture, a colored image 
was obtained on a white viewing screen. Consider the path of a cone of rays 
originating at the silver image in the emulsion layer of a lenticular film shown 
in highly exaggerated form in Fig. 66. The base is approximately 0.005 inch 
in thickness. On this, opposite to the lenticulations, is the emulsion layer 



THE LENTICULAR PROCESS 



227 



containing the silver images r, g, b, etc. These lie in the focal plane of the 
lenticular lens element M. The film is illuminated from behind by means 
of a beam of light with parallel rays. The aperture of the lens M, is the area 
covered by the image elements b, g and r. Since these are in the focal plane 
of M, the light passed by the silver deposits at b, g, and r, will be brought to 
a point focus. Let us concentrate our attention on the one ray from each 
of the silver deposits which passes through the nodal point of emergence A. 
This lies on the outer surface of the lens M, and therefore lies in the rear focal 
plane of the projection lens L. Consider the ray from the deposit b. After 
passing through A, it will continue in the same direction until it strikes the 
lens L. In a similar manner the rays from g and r will go through A and 
impinge on the lens L. As far as this lens is concerned, the three rays r, g, 




4, 



f££?*L> bcuse. 



FIG. 66 



and b originate from a point A in its rear focal plane, hence the lens L will 
project the three rays in the form of a beam of parallel rays. The direction 
of this beam is dependent upon the angle which the axis of the cone of rays, 
Abgr, makes with the optical axis of the lens L. Since it consists of parallel 
rays, the beam will come to a focus at the infinity plane of the lens L, which 
is usually 20 to 50 feet from its front nodal point. At this point, the three 
sets of rays originating at the points b, g and r in the emulsion layer of the 
lenticular film, will be fused into a single light point. The spectral charac- 
teristic, and the intensity of the light at this point, will be the sum of those 
of the three cones, symbolized by the rays r, g, and b. 

Now let us consider more carefully the individual cones of light. The 
intensity of the light beam originating at the point r in the emulsion layer 
of the lenticular film (Fig. 66), will depend upon the density of the silver 
image at that point. Thus the three cones r, g, and b, will have different 
intensities, each depending upon the density of the image at r, g, and b. The 
three cones unite at A, so that in so far as the lens L is concerned, it is being 
illuminated by means of a point: Source of light situated in the tear focal plane. 



228 HISTORY OF COLOR PHOTOGRAPHY 

But the lens M sees each cone from a different direction, hence will direct 
each cone in a different direction. Therefore, by the action of the lens M the 
three cones of light r, g, and b, become directed to three different portions 
of the aperture of the lens L. The banded filter in this aperture can be so 
arranged that the light in cone r passes through a red filter, that of g passes 
through a green filter, and that of b passes through a blue filter. 

If the lens L is equivalent to the lens used in the original photographing, 
the conditions for the exact reversal of paths will be satisfied, and the beam 
of light that passes through the filters in the aperture will be identical in com- 
position to the beam that gave rise to the deposits r, g, and b. As it leaves 
the lens L, the beam consists in reality of three sub-beams, each colored a 
different primary. But the three are brought to a focus at a point which 
corresponds to the infinity plane of the lens, so that here a blend will be 
formed. It is in this plane that the viewing screen is placed, or upon which 
the lens is brought to a focus. 

The short life of the lenticular processes was due to the failure of opticians 
to make lenses that approach the ideal discussed above. To a minor extent 
the emulsion makers were also at fault, as it was necessary to make fast emul- 
sions capable of resolving approximately 99 lines to the millimeter, if a film 
with 33 lenticules to the millimeter be used, and of a resolution of 75 lines 
to the millimeter if 25 lenticules be used. Fast emulsions, such as the East- 
man Super XX, have resolving power of approximately 45 to 55 lines if 
developed in DK-20, the fine-grain developer. Slower emulsions have higher 
values. Thus, the high-contrast Safety Positive 35 mm film will resolve 
upward of 70 lines to the millimeter, while the Microfile 35 mm film will 
resolve 135 lines. The negative materials that have the speed desirable for 
color work fall just short of proper resolution. Instead of getting distinct 
separations between the densities representing the three colors, they will be 
more or less overlapping and indistinct. 

It would take us too far afield to discuss at this point the theory of lens 
aberrations, or departures from ideal conditions and performance, but it is 
just these shortcomings that represent the biggest stumbling block in the 
path of successful utilization of lenticular processes. But we can discuss the 
effect produced, and indicate in that manner just what the nature of the 
problem is. The presence of a banded filter in the aperture of the lens means 
that the lens must be used at full opening, if only to allow the light rays from 
each filter area to reach each lenticular element. This means that the depth 
of field of the objective is very limited. It is true that each individual color is 
photographed only through one-third of the total aperture. Hence as far as 
that particular image is concerned it will have a somewhat greater depth of 
field than the sum total of the three colors. 

Consider a beam of rays entering a lens opened to full aperture, and coming 
to a focus in the image plane of the lens (Fig. 67). For convenience let us 
consider the beam to be parallel to the optical axis. Theoretically the entire 



THE LENTICULAR PROCESS 



229 



system of rays should come to a focus at the point M which is the rear focal 
point of the lens L. But this is far from the truth. In reality the different 
regions of the lens will bring the rays to a focus at different points, so that 

L 




instead of a center at which all the rays come together, there will be a definite 
area whose extent determines the circle of confusion. This is illustrated in 
Fig. 68, where it is seen that as the rays enter the lens further from the center, 

C' 




FIG. 68 

they are brought to a focus nearer to the lens. Also the image created by 
the rays is not a point, but a circle as indicated at A y B, and C. The best 
result from all practical considerations would be if the focal plane were posi- 
tioned at the point C. 







A^- ft 




^-^*r*L^^**^*^ 


* 1 r*^ ^*r 



FIG. 69 

In lenticular processes, this condition makes itself felt in this manner 
(Fig. 69). The rays A and B, will pass through the red filter area of the 
aperture. These being marginal rays, they will come to a focus at the point 
0, giving rise to a small dot with a definite area. The rays C, D y and E go 



230 HISTORY OF COLOR PHOTOGRAPHY 

through the central zone of the lens, where the green filter is positioned, and 
these rays will produce an image that does not coincide with 0, but lies beside 
it at N. A similar story is true for the blue rays. Lenses are usually corrected 
to make the total area of the three images smaller than /: 2000, which is just 
beyond eye resolution at normal viewing distance. A 10-inch lens would 
have a circle of confusion of approximately 1/200 inch. A 2-inch lens should 
have a resolution of 1/1000 inch, but the best of our 50 mm lenses have a 
circle of confusion that is distinctly larger, approximately 1/500 inch. If a 
film be used with more than 500 lenticules to the inch, and if a 2-inch lens is 
used, it may well be that the red cones will come to a focus on one lenticule, 
the green upon a second, and the blue upon a third. In that event, the three 
densities will not lie adjacent to each other, and will not be projected upon a 
common point. Color falsification results. It is true, of course, that the 
densities will not be far apart, but they will interfere with each other, for the 
direction of the blue ray that strikes a lenticule adjacent to the one for which 
it was intended, may coincide with the red or green rays that strike that 
lenticule. The density that results will be due to both colors, hence impure. 

The fact that the circle of confusion of a lens is so large when the focal 
length is increased, is the reason why no success whatsoever was ever attained 
with the lenticular process with film sizes greater than 35 mm. Even here 
the lenses were just barely usable, since the best 50 mm lenses gave images 
whose units of definition were only 1/500 inch. This could be much improved 
by working at a low aperture, but if three filters had to be positioned in the 
lens aperture, there was a minimum of effective diameter that could be used. 
Also the loss of light intensity occasioned by the use of not too efficient filters 
and by a three-way split of the beam, made it imperative that as large an open- 
ing as possible be utilized. 

This use of a very large opening makes itself felt in one other manner. It 
is well known that the depth of field is severely restricted when the lens is 
open wide. Hence the foreground is very sharp, but the background is quite 
diffuse, coming to a focus at an appreciable distance in front of the foreground. 
Thus the beam A, which originates in the foreground, comes to a focus at M 
in the image plane; but the beam B coming from the background, is focused 
well in front of the plane of M, at the point N (Fig. 70). By the time it has 
reached the focal plane, the rays have spread on into a cone N BGR, where 
B, G, and R represent the blue, green, and red sections respectively. This 
cone may cover an area on the film which includes more than one lenticule, 
and therefore color distortion and degradation will take place. If to this is 
added background motion, then fringing is bound to occur, since the three 
colors will become imaged at non-adjacent portions of the lenticular film, 
and will therefore not re-combine upon projection. 

These are not the only optical difficulties that make lenticular processes 
difficult to handle, but they are typical of the others. It would take us too 
far afield to discuss all of them, and it would do very little good, since the 



THE LENTICULAR PROCESS 



231 



problem of aberrations has now been known and discussed for almost fifty 
years without solution. Lenses have been tremendously improved by making 
corrections to minimize the effects of aberrations, but all efforts to completely 
remove them have so far been futile. Until that time, or at least until the 
diffusion caused by the aberrations in a lens operating at full aperture (/: 1.9), 
becomes considerably less than 1/500 of an inch, and until lenses can be made 
with a depth of focus at/: 1.9 equal to that at 7:32, lenticular processes will 
not be wholly satisfactory. 




FIG. 70 



These difficulties arise only when lenticular film is used for the original 
exposure. Up to the advent of the monopacks, this appeared to be the de- 
sirable technique, for the inconveniences of one-shot cameras, their bulk, 
delicacy, inefficiency, etc., made an integral process very desirable. This 
condition was rectified to a great extent by the introduction of Kodachrome 
and similar materials. The best position lenticular film can occupy would 
seem to be in the making of positive prints for projection purposes, from separa- 
tions made in some other manner. It is possible to make separations of high 
quality from Kodachromes if a masking technique is used. These can be 
printed upon lenticular film, through lenses that are exactly equivalent to the 
projection lenses. This proposal has not received a very favorable reception, 
although the idea has been suggested many times. It will be discussed more 
fully in Chapter 18. 



CHAPTER 17 
LENTICULAR DISCLOSURES 



Th, 



.HE patents dealing with lenticular processes can be classified into a 
number of sections, thus: 

i. The Preparation of the Film Base. 

2. The Properties of the Film. 

3. The Lenses for the Taking System. 

4. The Diaphragm and the Filters. 

5. Projection Systems. 

6. Duplicates by Projection Printing. 

7. Duplicates by Contact Printing. 

8. Separations from Lenticular Film. 

9. Printing Separations upon Lenticular Film. 
10. Other Uses of Lenticular Film. 

Of these, sections 1 to 4 deal with the raw material and auxiliary apparatus. 
These patents cover approximately one-third of the total number. The 
remaining sections, 5 to 10 inclusive, deal with the uses to which the film 
could be put. More than 65 per cent of the patents are distributed in these 
groups. Fully one-third of these deal with problems that arise in the duplica- 
tion of the lenticular positive, by projection printing. Some idea of the 
popularity and esteem which this process had can be obtained when it is 
realized that in the ten years from 1928 to 1938, more than 250 patents deal- 
ing with lenticular processes, were issued in England. Because patents are 
issued later in the United States than in most other countries, not nearly 
that many were issued here. Evidently the sharp decline in interest in this 
method, coincident with the development of Kodachrome, Ansco Color, and 
color-coupling in general, resulted in the lapsing of many patent applications. 

Although R. Berthon was the first to disclose the possibility of color repro- 
duction by use of lenticular film, he did not disclose how he would prepare 
such a material. For this, we are indebted to A. Keller-Dorian, another 
Frenchman, who saw the tremendous possibilities of lenticular film, and who 
thereafter devoted most of his attention to its preparation and improvement. 
He proposed to pass celluloid sheets between two rollers (Eng. P. 24698/14; 
Fr. P. 466781; U.S.P. 1214552). One of them was smooth, the other was 
engraved so that one side of the celluloid became embossed. The embossing 
could be in the form of hexagonal pyramids (Eng. P. 7540/15, 207836 and 
207837; Fr. P. 472419) or hemispheres. The interstices between elements 

232 



LENTICULAR DISCLOSURES 233 

were to be filled in with black (Fr. P. 516050), a procedure that could be ac- 
complished photographically by the use of a special diaphragm (Fr. P. 547529, , 
and 54753°; En g- p - 130656). 

The machine for embossing the film surface was the subject matter of many 
patents. J. Audibert (U.S.P. 1625586) used highly polished rollers. A. Os- 
wald (U.S.P. 1855198; Eng. P. 395200) arranged to regulate the tension and 
temperature at those sections of the rollers that were in contact with the film. 
The same company, working under the Keller-Dorian original disclosures, 
disclosed other types of machines to accomplish this purpose (Eng. P. 298242, 
298951, and 389830). 

The Eastman Kodak Company was also interested. P. R. Ord, of its staff, 
proposed to make a model embossing roller, by winding a wire upon a tem- 
porary core, (U.S.P. 1757543). This was plated, the core and wire were then 
removed, and the hollow left by the removal was filled in. H. E. Hastings 
(U.S.P. 1813669; Eng. P. 330151) proposed to make an embossing tool out 
of extra-hard drawn phosphor bronze. This was polished, then grooved with 
a series of closely spaced serrations. The grooves were polished, and finally 
chromium plated. O. Wittel, also proposed to use wires (U.S.P. 1880632; 
Eng. P. 329214). But instead of using them as a model upon which to electro- 
plate a shell, he used them as a tool to cut a spaced series of grooves into a 
metal cylinder that was to be used as the embossing roller. 

Another method for cutting grooves in an embossing cylinder, was disclosed 
by P. Brosse (U.S.P. 1703026; Eng. P. 265970). He used a diamond cutting 
tool, perfectly shaped, to cut through a gold or silver plating on a roller. 
The Agfa Ansco Corp. and the I.G., would also use a wire coiled around 
a cylinder as the cutting tool (U.S.P. 1879237; Eng. P. 341948, 375225). Or 
a cast from this could be used. In another disclosure, they proposed to use 
dichromated gelatin reliefs in the exact shape of the lenticules, with which 
to emboss the celluloid sheet (Eng. P. 433875). 

The Siemens and Halske Company, of Germany, was a prolific patentee in 
all phases of the lenticular processes. The preparation of an embossing ma- 
chine was disclosed in a series of patents (U.S.P. 1945586, 1945935, 1992279, 
and 1994054). They also were inclined to use wire with which to cut grooves 
in the embossing cylinders. Other types of machines were disclosed by 
A. Rodde, A. H. Herault, V. Hudeley, and J. Lagrave (U.S.P. 1955658; Eng. P. 
406163), C. Roehrich (U.S.P. 2048816), and P. Fournier (Eng. P. 244740). 

The shape and size of the lenticules were quite important, and it is no wonder 
that a number of patents were issued to regulate them. The Keller-Dorian 
disclosures have already been discussed. Berthon in his original patent did 
not dwell long upon this phase, but in later patents he disclosed the use of 
cylindrical lenticules (Fr. P. 402507) or prismatic surfaces (Fr. P. 401342). 
He discussed the problem of number in a separate disclosure (Eng. P. 264123; 
U.S.P. 1707157). Two conditions must be fulfilled. First, the lenticule 
must be optically perfect. Second, the diameter of the lenticule should be 



234 



HISTORY OF COLOR PHOTOGRAPHY 



an accurate function of the relative aperture of the camera, the thickness 
of the film base, and the curvature of the lenticules. Thus with an objective 
operating at an opening of /: 2.5, a film base with a thickness of 120 to 130 /x, 
and a radius of curvature for the lenticules of 0.04 mm, the diameter of the 
elements should be 0.04 mm. This is equivalent to 22 to 23 elements per 
millimeter. But the limit of definition of an image should correspond to 
50 elements per millimeter. This is possible only if the effective thickness 
of the base is reduced. This can be accomplished, without sacrificing any 



backincf 




tker 



jfu/e 
fn *s,Qj*stuue. emulsion* 



FIG. 



B 

71 



film strength, by the following procedure. The thin lenticular base is coated 
on its lenticulated side, with a layer of transparent material whose index of 
refraction is considerably different from that of the film base. Thus in Pig. 71 
a lenticulated celluloid surface C, has an emulsion B, coated on the side op- 
posite the lenticulations, and has a layer of bakelite coated over the lenticulated 
surface. This is a material with a much higher index of refraction than cellu- 
loid. It is also possible to use for an overcoating, a material whose index of 
refraction is lower than that of celluloid, in which case the curvature of the 
lenticules would be in the opposite direction, as is illustrated in Fig. 72. 



£Lcktn(f 




FIG. 72 

The Keller-Dorian Company proposed to use differently shaped lenticules 
(Eng. P. 246829). In other disclosures, it was proposed to emboss more than 
25 elements to the millimeter (Eng. P. 207836, and 207837), and to make the 
width of each lenticule from 10 to 40 per cent greater than the width of the 
image of the filter bands projected by each element on the sensitive surface 
(Eng. P. 261363). This would insure that the silver deposits would be distinct, 
with no overlapping of the images behind different lenticules. Before emboss- 
ing, the film was partially dried. The exact opposite, treatment of the film 
base with swelling agents Was proposed by G. Heymer (U.S.P. 1996868] 
Eng. P. 375229)- 



LENTICULAR DISCLOSURES 



235 



The use of a sensitive material other than silver halides, has also been 
proposed. R. Berthon (Eng. P. 274837) suggested diazotype, with gas de- 
velopment. In another disclosure it was proposed to use dichromated gelatin 
or diazonium salts (Eng. P. 289864). A feature of this last is that the 
lenticulations are present in the interface between the celluloid and the 
dichromated gelatin, hence the two outer surfaces are plane and parallel to 
each other (Fig. 73). A film of this type would have some advantage in pro- 
jection as it would not be subject to scratching nearly as much as in the case 
where the lenticules are on one of the outer surfaces. 



oeUcciotd 




cCCcAro 9 m, ate.cC 



FIG. 73 



The lenticular surface in conjunction with a proper lens system is the optical 
analogue of the screen plate. Therefore the objective and the filter system 
used with it, become a very important matter. Nor was this subject disre- 
garded by the experimenters. The conditions were outlined generally, in the 
basic patents of R. Berthon (Eng. P. 10611/09; Ger. P. 223236; U.S.P. 
992151; Fr. P. 399762 addition 11 286, 413 103 addition 12342, 430600 addi- 
tions 14438 and 14439). Here was disclosed the use of a symmetrical lens 





'fetter* 

FIG. 74 

system in whose center was positioned the set of banded filters. A beam of 
parallel rays entering the lens in the direction of the arrow (Fig. 74), will be 
divided into three beams each colored in a different primary, by the banded 
filter situated in the center of the lens system. The lens will convert each 
of these colored beams into cones whose apices coincide at the nodal point 
of admission of each lenticular element. Since, as far as this elementary lens 
is concerned, the three cones are not parallel, three images in juxtaposition 
will be formed on the photographic emulsion, on the opposite side of the 
lenticular surface. 



236 



HISTORY OF COLOR PHOTOGRAPHY 



Some of the intricacies involved in lens construction are outlined in the 
patent disclosures of I. Kitroser, of the Keller-Dorian Company (Eng. P. 
348465 and 384009; U.S.P. 1897262). As shown in Figs. 75 and 76, he pro- 

N 







FIG. 75 



posed to place a negative lens iV, well in front of the objective 0. In front 
of the lenticular film he placed a collimating lens C. The focal length of C is 
50 mm, which is the same as that of 0. The focal length of the negative lens 
will be — /n, and it is placed a distance e in front of the nodal point of admission 




FIG. 76 

of the lens 0. If the collimating lens C is made of extra heavy glass with a 
refractive index of 1.8, the curvature of the field (Petzval curvation) intro- 
duced by it will be given by 



D = - 



fXn 50X1.8 



90 



This means that the field will have a radius of curvature of 90 mm. In the 
corner of an 18 by 24 mm field that is covered by the objective at an opening 
of 7:2.5, the added circle of confusion will be approximately 0.5 mm. Now 
let the distance e be 100 mm, let/u be 200 mm, and assume n equal to 1.50. 
The combined focal length of the lens N in combination with 0, will be: 



F = 



/X/n __ 50 X (— 200) — 10000 

fn+f-e -200 + 50-100 - 250 



= 40 mm 



LENTICULAR DISCLOSURES 237 

The objective 0, will be at a distance x from the film, such that : 

1 1 1 1,1 1 

- = t—j- + -1 = 1 = — or a = 60 

The collimating lens will need to have a focal length of 60 mm. The Petzval 
curvature of the combination will be the sum of the curvatures of the undivided 
units, and this will be: 



'--*&)" 



60X1.8 200X1.5 168.7 



or the radius of curvature will be 168.7 nun. This will give a circle of con- 
fusion at the edges of an 18 X 24 mm field that is equal to 0.25 mm. 

The effect of placing a negative lens well in front of the objective is there- 
fore quite evident. Ordinarily, a lens with a focal length of 50 mm will be 
situated so that 50 mm separate the film and the nodal point of emission of 
the lens. By placing a negative lens with a focal length four times that of 
the objective, a distance equal to half of its own focal length in front of the 
objective, the focal length of the combination becomes decreased to 40 mm. 
At the same time the film is to be placed a distance of 60 mm behind the 
nodal point of emission of the objective. This is a fact well worth remem- 
bering. Technicolor, who uses a solid glass prism to split the light beam with 
two sub-beams, made use of this property to eke out some free space be- 
tween film and prism. 

One other effect of such a system is worth noting, and that is the flattening 
effect it has upon the Petzval curvation. This, it must be remembered, arises 
from the fact that the oblique rays come to a focus somewhat in front of the 
rear focal plane. Thus in Fig. 77, the rays A which are parallel to the optical 



A A- 




'T 


^^^ >o j A — 1 — ^^ 




^a^AjLI 


/ 
/ 

/ 
/ 


R 







FIG. 77 

axis, are brought to a focus at i?, in the optical plane. But the oblique rays B, 
are brought to a focus at the point S, somewhat in front of the plane RT. 
By the time the rays reach the focal plane they are spread out, and instead 
of being a point, the image becomes a circle. If a line were drawn through 
all the points S, which represent the true foci of the light beams entering the 
lens, an arc would be formed. The radius of this arc is the Petzval curvature. 
It is determined by the focal length and the index of refraction of the lens, and 



238 



HISTORY OF COLOR PHOTOGRAPHY 



is independent of all other constants. The greater the value of this radius 
of curvature, the flatter the image plane, and the smaller the circle of confusion 
for the uncorrected lens system. By use of a negative lens, this has been 
reduced from a value of 0.50 mm to 0.25. 

It is a simple matter to determine what the circle of confusion would be in 
terms of the Petzval curvature. Let us take the case discussed above (Fig. 78), 




FIG. 78 

where the Petzval curvature was equal to 90 mm. Then the distance OA 
= OC is 90 mm. Let the point R be the optical center of the lens system so 
that RC = 50 mm. Since the point A is in one corner of the frame, which 
is 18 by 24 millimeters in size, the distance AB will be one-half of the diagonal, 
or 15 millimeters. The angle at D is then one whose sine is equal to 15/90 
or 0.1667. Therefore the cosine would be equal to 0.986. The distance OB 
will be given by 90 X 0.986 or 88.74. Hence the distance BC will be equal to 
1.26 millimeters. 

Consider now the case of a beam of light A (Fig. 79), entering a lens system 
L, operating at an aperture of 7:2.5, and giving rise to a Petzval curvature 




at the point A equal to 90 mm. Then the distance PD will be 1.26 mm. 
In so far as the image BC of the rays A is concerned, the point A' will act as 
a pinhole camera, and will yield an image in the plane through D, of the 
aperture of the lens L. Therefore the size of the diameter BC of the circle 
of confusion, will be related to the aperture of lens L as the focal length of 
the lens A 1 is related to the focal length of the lens L. The focal length of the 
lens A' is 1.26 mm and of L it is 50 mm, so that the size BC is 1.26/50 of 



LENTICULAR DISCLOSURES 239 

the size of the aperture of the objective, and since this is operating at 7:2.5, 
it will have an aperture that is 20 mm in diameter. Hence the circle of confu- 
sion will have a dimension of 



20 X 



(= -^— ) or 0.504 millimeter 
50/ 



5o 

The introduction of the collimating lens C, into the lenticular system 
(Figs. 75 and 76), caused the formation of a disc of confusion equal to 0.5 mm. 
But placing a negative lens well in front of the objective cut this down to half, 
to a value of 0.25 mm. This value for the size of the image dot gives some 
realization of the difficulties encountered with lenticular processes. There 
are approximately 600 lenticules to the square mm, hence in an area 0.25 mm 
in diameter there will be approximately 30 lenticules. But for really accurate 
results, the image must fall on only one lenticule. That is why out-of-focus 
and marginal images are very poorly portrayed in lenticular systems. 

The best lenses used in miniature cameras have a disc of confusion equal 
to 1/500 inch, or 0.05 mm. Lenses with longer focal lengths than two inches 
will operate with a disc of confusion that is more nearly 1/200 inch. Lenses 
of longer focal length operate at a disc equal to 1/200 inch. Since the size 
of a lenticule is approximately 0.04 mm or 0.0016 inch, it is seen that even 
the best of our lenses will give an image that is much coarser than the individual 
lenticule. The use of longer focal length lenses becomes highly undesirable 
in lenticular processes because the image of a single beam will encompass 
many more than one lenticule. 

Other lens systems were disclosed by G. Grosset, V. Hudeley, and J. La- 
grave (U.S.P. 2003881; Eng. P. 41x322), J. G. Capstaff (U.S.P. 2091699) and 
C. Nordmann (Eng. P. 411024, 424906, and 425058). The Capstaff and the 
last of the Nordmann patents are interesting in that they disclose the use of 
a light splitter behind the lens system. This splits the light into three sub- 
beams that are parallel to each other, and which first pass through a set of 
filters, then through another lens which brings all three to a common focus. 
This is illustrated in Fig. 80. The light beam R enters the objective 0, where 
it is split into three parallel beams A, B, and C, by the light-splitting prism P. 
These go through the niters F, each sub-beam going through a differently 
colored section. From here the three parallel beams go to the lens L. Since 
this sees the three as parallel rays, it brings them to a point focus at S, the 
plane containing the lenticular film. The advantage which this procedure 
has over the normal one is that each sub-beam contains the rays from the 
entire aperture of the taking objective, hence can be modulated by means 
of an iris diaphragm, to regulate the light intensities. The two lenses must 
be carefully adjusted, however, to allow critical focusing for objects in front 
of infinity. 

The original Berthon patent utilized a circular diaphragm, with the three 



240 



HISTORY OF COLOR PHOTOGRAPHY 



filters completely filling the aperture. This was placed in the center of a 
symmetrical lens system (Fig. 74). Such a system created a problem of vary- 
ing the lens opening in such a manner that the intensities of the three cones 
leaving the lens system became affected in a like manner. Obviously an iris 




FIG. 80 






diaphragm, operating by changing the effective diameter of the lens aperture, 
would affect the two outer zones much more than the central one, hence 
some other scheme must be adopted. This is strictly a problem in mechanical 
design. The most popular solution, patented by almost every worker dealing 
with this problem, was to make a rectangular diaphragm. This and other 
schemes are disclosed in the series of patents noted below: 

U.S. Patent 

1687055 J. G. Capstaff, to Eastman Kodak Company 

1688370 0. Wittel, to Eastman Kodak Company 

1688441 D. H. Stewart, to Eastman Kodak Company 

1689258 M. W. Seymour, to Eastman Kodak Company 

1 7083 7 1 M. W. Seymour, to Eastman Kodak Company 

1754282 F. H. Owens, not assigned 

1 762 143 J. G. Capstaff, to Eastman Kodak Company 

1762938 D. H. Stewart, to Eastman Kodak Company 

1 769041 M. W. Seymour, to Eastman Kodak Company 

l8 73758 A. J. Ginsberg, not assigned 

191 2700 E. C. Fritts, to Eastman Kodak Company 

1944230 C. L. Gregory, Kislyn Corporation 

1949414 C. L. Gregory, Kislyn Corporation 

1989 134 F. Fischer, Siemens and Halske 

1989748 H. Frieser, Siemens and Halske 

2029614 F. Fischer, F. Strecker, H. Neugebauer; Siemens & Halske 

2041465 N. Gehrke & G. Strohmenger, to I.G. 



LENTICULAR DISCLOSURES 241 



English 




287488 


Keller-Dorian Co. 


295995 


Kodak Ltd. 


33i97i 


Keller-Dorian Co, 


340881 


Kodak Ltd. 


350481 


I.G. 


356977 


Kodak Ltd. 


366175 


I.G. 


368970 


I.G. 


373861 


Kodak Ltd. 


387159 


I.G. 


401963 


I.G. 


401982 


I.G. 


406187 


I.G. 


406304 


E. Leitz 


440582 


Opticolor 


44i 7 13 


Opticolor 


454357 


Opticolor 


459634 


Opticolor 


492850 


Opticolor 



The use of gelatin niters was open to some objections, in that fading may 
occur. I. Kitroser would overcome this by the use of colored glass (Eng. P. 
385680; U.S.P. 1878857), ground to the proper thickness to give correct 
spectral transmission. The three glass thicknesses were evened out by adding 
uncolored glass to each. In some other patents (Eng. P. 286223 and 303170) 
it was proposed to use liquid niters in tubes placed with 'their axes parallel 
to the lenticulations. 

Instead of using three bands (red, green, and blue), it was proposed to 
place a yellow band on one side of the set, and an orange one on the other 
(Eng. P. 294493). These were to act as compensators for the deficiencies of 
the three-color system. It is open to grave doubt whether any improvement 
would result from the use of the added yellow and orange bands. These would 
first of all divide into five, instead of three, the area covered by each lenticule, 
and this already requires an emulsion whose resolving power is approximately 
1500 lines to the inch. To operate with five bands instead of three will require 
emulsions whose resolving powers would be approximately 2500 to the inch, 
and only the slower emulsions would have this resolution. Then again, the 
densities behind the orange and yellow lines, unless these colors be spectral 
yellow and spectral orange, would be the same as those behind the red and 
green, so that color degradation would result. In English patent 314995, 
the same company proposes to use four bands instead of three, with the two 
outer bands having the same color. The I.G. Company proposed to move 
the filters after each exposure, the distance of one band, but kept the colors 
in the same cyclic rotation (Eng. P. 449762). This meant that the order red, 
green, blue, red, green, blue, etc. would be maintained, but if the central band 



242 



HISTORY OF COLOR PHOTOGRAPHY 



were green in the first exposure, it would be red in the second, and blue in 
the third. 

Instead of putting the filters in the optical center of a symmetrical lens, 
the Keller-Dorian Company proposed to put them well in front of the lens 
system. .(Eng. P. 293047 and 331971). This was also proposed by B. E. 
Luboshez (Eng. P. 3395 n) who claimed that by this procedure the image 
of any one beam will cover more than one lenticule, thus smoothing out the 
final image. 

L. Tissier would do away completely with the banded filters in the lens 
aperture. Instead, a special lens is placed in front of the lenticular film, 




oDjec&)*e.o 



FIG. 81 



which causes dispersion to take place (Eng. P. 24276/14). In order to insure 
that no rays reach this lens in a direction perpendicular to it, the central 
region of the camera objective is blocked out. Thus (Fig. 81) the light beam 
entering the camera objective 0, is broken up into oblique rays by the dia- 
phragm D in the lens aperture. The oblique rays then strike lens C, which 
is made of two components with widely differing indices of refraction. In 
effect, lens C is merely a direct-vision prism which disperses the rays incident 
upon it into their respective spectra. The spectrum of each beam is thus 
imaged individually by the lenticules. The use of the diaphragm D results 
in a considerable loss of intensity. It can be replaced by a lens situated di- 
rectly behind the lens C, and having its optical axis different from that of 
and C. 

Another scheme to replace the filters was outlined by A. Keller-Dorian 
(Eng. P. 158511; Fr. P. 523336; U.S.P. 1372515). Behind the camera 
objective 0, is placed a reflecting prism P, and adjacent to the reflecting sur- 
face of this prism, is placed a glass plate R, slightly separated from the prism 
at one end. The separation gives rise to interference bands, whose colors 
take the place of the niters (Fig. 82). 

The rigid registry requirements of lenticular processes make the processing 
of the film by reversal the natural method for its treatment. Where but one 
copy is desired, as in amateur 16 mm motion pictures, lenticular processes 
do not create many problems. The curvature of the field, however, creates 
one such problem, and M. W. Seymour proposed to overcome this by making 



LENTICULAR DISCLOSURES 



243 



the gate of both the camera and the projector in the shape of a curve (Eng. P. 
318040). The great loss of light occasioned by the use of gelatin or glass 
niters was minimized by the LG. by the use of direct-vision prisms in front 
of the gate of the projection machine (Eng. P. 383795). This could be kept 
in constant motion (Eng. P. 446282 and 447490). Instead of prisms, it was 
suggested that gas-discharge lamps be used. Three are needed, and they are 
so positioned in the projector lamphouse that their images fill completely the 
areas that would be occupied by the filters (Eng.. P. 419701). Projectors use 
lamps that give out great heat which may affect not only the film, but also 
the filters. The use of gas-discharge lamps reduces the heat tremendously. 
The colors so formed can be controlled by physical means to conform to the 




'/enticu/ar ftt#t 



FIG. 82 



three primaries, and the intensities would be much greater than those pro- 
duced by means of a single light source behind gelatin filters. As far as the 
projection lens is concerned, it would see only three bright colored areas in 
the position that would ordinarily be occupied by a set of banded filters. 
These also would be seen by the projection lens as three colored areas, so the 
one system is identical to the other. The question of heat was solved by 
Opticolor by inserting a cooling coil in the lens and filter areas (Eng. P. 432401). 
The projection of a film positive is usually carried out by means of ldnses 
whose optical characteristics are different from those of the taking lenses. 
This creates quite a few optical problems. In order to recreate the original 
scene, the light that passes through any given dot in the film must be directed 
to the same filter that the original beam passed through in giving rise to this 
dot image. The exact conditions that must be fulfilled by projection lenses 
are described by I. Kitroser (Fr. P. 537508). Expressed briefly, these are 
that the filter, when viewed from any single lenticule, must subtend the same 
angle in the projection system that it subtended in the taking system. A 
somewhat more detailed discussion is contained in a disclosure by A. Avexan 
and I. Kitroser (Eng. P. 308320). 



244 HISTORY OF COLOR PHOTOGRAPHY 

For projecting lenticular film, the filter bands are placed in the lens dia- 
phragm. Since the focal length of the projection system is considerably 
longer than that of the taking, this would mean that the filter bands are 
positioned a greater distance in front of the gauss points in the projection than 
in the taking system. It becomes necessary therefore, to find some means of 
bringing these points up nearer to the niters. This can be done by the use 
of a supplementary lens, placed in front of the projection lens. The new 
optical system must fulfill the following conditions: 

i. The diaphragm must coincide with the outgoing aperture, and must be 
positioned in front of the objective, at a short distance from the front 
optical element. 

2. The filter band position must be made to coincide with this diaphragm. 

3. The new system must move the position of the film so that it will be the 
same optical distance behind the projection diaphragm that it was 
behind the taking diaphragm. 

4. The opening of the diaphragm must correspond to the cross section of 
the beam projected by each lenticule upon the lens aperture. 

In designing such lens systems, use is made of the fundamental law of optics 
governing the coupling of lenses. This is expressed mathematically as fol- 
lows: 

J h+h-e. 

Here / is the focal length of the coupled system, /1 and fo the focal lengths of 
the individual lenses, and e the distance from the nodal point of emission of 
the first lens to the nodal point of admission of the second. From this ex- 
pression it is seen that the focal length of a combination of lenses becomes 
equal to that of any one of the components if e is equal to the focal length 
of that lens. Thus if e = /1, then / = /1. If e is made less than the smaller 
of the two, then/ becomes smaller than either, /1 or / 2 , If e is larger than the 
greater, then/ is greater than either /1 or/2. If e is made larger than/i +/ 2 , 
then the character of the lens is changed, a divergent system becomes con- 
vergent, and a convergent system becomes divergent. The use of supple- 
mentary lenses for this purpose is disclosed by C. W. Frederick (Eng. P. 
294207, 330810; U.S.P. 1685600 and 1749278); P. Rehlander (U.S.P. 
1872501); C. Nordmann (Eng. P. 360524); and the I.G. (Eng. P. 342924). 
Other systems have been described by Pathe Freres (Eng. P. 17330/13), 
A. Dvornik (Eng. P. 421120), and the I.G. (Eng. P. 436871 and 447834). 

Of especial interest are the disclosures of Frederick, in that he reviews the 
optics of lenticular projection. The object which he set up for himself was to 
determine the position of the filters in a taking system, so that the images 
of the filters, as viewed from any point in the plane of the lenticulations, will 
appear to coincide in size and in shape with the filters in the projection system 



LENTICULAR DISCLOSURES 



245 



when viewed similarly. Every lens system has two gauss points, G' and G" 
(Fig. 83). A ray AG', proceeding from the object to one of these, willappear 
to proceed in a parallel direction G"A', from the other. There are two other 
points of importance. Light, coming from infinity and entering the lens 
system from either side, will be brought to point foci. These are the front and 
rear focal points, and 0', of the system. If G' and G" are the front and rear 
gauss points respectively, then the distance OG' must necessarily be equal 
to the distance O'G". 



-~^c 



0<t: 



n! in' r 
>-*t"f—- 
I ' 

1 1 



:=^-r 



4 a' 
0' 



A-*" 



-. 



Si 



filf 1 
1 1 

FIG. 83 



8' 



Suppose that the lens has an aperture equal to/: 2, and that HL represents 
the diameter of this aperture, which is situated at the front nodal point G'. 
From any point in the, rear focal plane A'O'B', this will appear to be posi- 
tioned at G". The aperture will subtend an apparent angle H'O'IS, with the 
axis. Stated another way, an object placed at G r will appear to be placed at 
G" when viewed from the focal plane where the lenticulations are placed. 
If the object LH be moved nearer to 0, its vertical image will appear in a 
plane that is determined by the expression: 

where / is the focal length of the lens system, /1 the distance of LH from G', 
(object distance), and / 2 the distance of the image CD' from & (the image 
distance). This expression is the fundamental relationship in lens optics. 
If the object occupies the plane CID, where IG' is less than half of the focal 
length of the lens (less than half of OG'), then the virtual image, C'D' y will 
occupy a plane somewhere between and 7. When I moves close to G', 
its image V is but a slightly larger distance from G". As I moves away from 
G' to the left, its image will move away from G" at a slightly faster pace, also 



246 HISTORY OF COLOR PHOTOGRAPHY 

to the left. When I reaches 0, then the object distance becomes zero, and the 
image distance becomes infinite, so that the image becomes indeterminate. 
Thus a beam originating at a focal point of a lens system, will project a beam 
of parallel rays whose focus is at infinity, and vice versa. When I passes to 
the left of 0, it changes its sign (/i changes from a positive to a negative 
number, or vice versa depending upon how the co-ordinate system is chosen). 
Since the focal length, /, of the lens is always a positive number, this means 
that the sign of / 2 must change, and instead of giving rise to a virtual image 
to the left of G", there will be formed a real image to the right of the point. 

Although the size and position of an object placed at CID, somewhere be- 
tween and G', can be determined from the basic lens equations, it is also 
possible to do so graphically. The rays that proceed from the focal point 
of the lens system, emerge as parallel rays. We assume that the position 
CID, and the points 0, 0'G f , and G" are known. Along a line which represents 
the optical axis, mark off the points 0, 0', G', and G". OG' must be equal to 
O'G", and this is equal to the focal length of the lens system. Now mark 
off the line CID, to represent the position of the object whose image is de- 
sired. Draw lines through G 1 and G" to represent the nodal planes. From 
0, draw the lines OC and OD. As far as the lens is concerned, these are rays 
which originate at its focal point, so that they will proceed from the lens in 
the form of parallel rays. Since they originate on the optic axis, the emergent 
rays will be parallel to the optic axis. The rays intersect the nodal plane of 
admission (the plane through G f ) at the points N and M. When viewed 
from any point in the rear focal plane A'O'B', these points of intersection will 
appear to be the points N', and M' in the plane through G" (the nodal plane 
of emission). Therefore, through N' and M' draw lines parallel to the optic 
axis, to represent the emergent rays. These are the lines N'R and M'S. 
The images of the points C and D must lie somewhere along these lines. 

Another well-known rule of optics is that parallel rays entering a lens system 
will come to a focus somewhere in the rear focal plane. If these rays are 
parallel to the optic axis, then the point focus coincides with the rear focal 
point. From C and D, draw lines parallel to the optic axis. These intersect 
the front nodal plane at L and H. From any point in the rear focal plane, 
they will appear to be positioned at L' and H f , in the nodal plane of emergence. 
As far as the lens is concerned, the lines CL and DH represent parallel rays, 
which must come to a focus at the point 0', and appear as if they emerged 
from the points L' and H'. Therefore we draw the rays L'O' and H r 0' and 
along these rays must be the images of the points C and D. Extend the lines 
O'U and O'H' until they intersect the lines N'R and M'S, which also con- 
tain the images of C and Z>. The intersections of these lines, C and D', are 
the images of C and D respectively. Therefore, the object CID, placed some- 
where between the front nodal point G' and the front focal point D, will 
appear, when viewed from a point in the rear focal plane, as if it were posi- 
tioned at C'l'D'. 



LENTICULAR DISCLOSURES 247 

It is not always possible, in designing lens systems, nor is it always desirable, 
to place the niters in the front nodal plane of the lens system. The best posi- 
tion for them is in a plane that is just in front of the front surface of the 
lens system, such as the position CID in Fig. 83. In so far as the lenticular 
elements situated in the rear focal plane A'O'B' are concerned, the filters 
will appear to occupy the position C'I'D' y and they will subtend the angle 
C'O'D' with the point 0' in the rear focal plane. The conditions laid down by 
Kitroser are that the image C'I'D' must subtend the same angle with each 
lenticule during projection that was subtended during the original taking. 
Then, and only then, will the densities formed by the red portion of the beam 
become directed through the red section of the filter. If the same lens system 
is used for viewing as for taking, no difficulties arise, but if different lens 
systems are used, then complications begin to make themselves felt. 

It has been demonstrated that the apparent size and position of an object 
can be varied at will, merely by placing it somewhere between the points 
and G in the lens system. Let us suppose that we are forced to use a projection 
system, such as was described in Fig. 83, and that the filters occupy the posi- 
tion CID. Let us suppose that we are likewise forced to use a taking system 
whose focal length is one-half that of the projection system, i.e., og' = o'g" 
= \ O'G" * The effective apertures of the two lenses are the same, /: 2 so that 

HL/O'G" = -^7 and hi = £ o'g", EL = i O'G". 

The question is where to place the filters in the taking system so that the 
images of the filters will subtend the same angle in both cases, with correspond- 
ing points in the rear focal planes. 

On Figs. 8^ and 84 the following data are known, the points 0, 0', g' and g" 
and the angle C'O'D'. If we can position the niters c, i, d, in the taking system, 
in such a manner that the image will appear at c'i'd', where i'o' = I'O', and 
c'd' — CD' , then the problem is solved. To this end draw the line 00' to 
represent the optic axis. Along this line, mark off the known points g' and 
g", and then mark off the point i' so that i'o' = I'O'. At the point i', draw 
a line perpendicular to the optic axis, and on this line mark off the points c' 
and d' so that c'i'd' becomes equal to C'I'D'. We desire now to find the 
position and size of an object, to be placed somewhere between g' and 0', 
so that its apparent image, when viewed from the point o l , will coincide with 
c'i'd'. From c' and d' draw lines parallel to the optic axis. These intersect 
the frontal plane at the points n and m. From these points draw the lines on, 
and om. The capital letters refer to the projection system (Fig. 83), while the 
lower case letters refer to the taking system (Fig. 84). 

Somewhere along these lines will lie our object points. From c' and d' 
draw the lines c'o' and i'o 9 . These will intersect the rear nodal plane at the 
points V and h'. From these points draw lines parallel to the optic axis. These 
will intersect the lines on and om at c and d, which are the desired points. If 



248 



HISTORY OF COLOR PHOTOGRAPHY 



*- 



SL4- 



— ->r 
• i 

J I 

A-&- 



A 



—V 



.;*-T 



z~-^K? 



I I 



rfU-J 






7HS 



FIG. 84 



the filters are placed in the position occupied by the object cid, then from 
any point in the rear focal plane, such as from the point o', they will appear 
to be positioned at c'i'&\ and to subtend an angle c'o'd* with the lenticule 
at o'. This is identical to the view had by the lenticule at 0, in the projection 
system. Hence the two systems are exactly equivalent. In his patents (cf . 
above) Dr. Frederick gives the complete data for the construction of lenses 
complying with these conditions. 



CHAPTER 18 
DUPLICATION OF LENTICULAR FILM 



Th: 



.HE image in lenticular processes consists of a honeycombed series of 
densities. In each unit there lie three juxtaposed elements, each of which 
constitutes the density of a given primary color. Each unit lies adjacent to 
and touches the neighboring units. It is an easy matter to duplicate the 
values for the densities at each point in the image space, but that is hardly 
sufficient for the purposes of forming duplicates of lenticular images upon 
lenticular film. It is necessary to duplicate the exact registry of the individual 
densities with respect to the lenticules, so that they will subtend the same 
angle with the nodal points of admission of each lenticular lens in both master 
and copy. This makes the duplication of the film a rather complicated optical 
problem. Unless this registry is carefully maintained, poor color separation 
results. When this problem has been overcome, the technician is confronted 
by the problem of moire (arising when the pattern of the master film is placed 
on top of the pattern of the copy film), and the problem of wedging, which 
is caused by the fact that the marginal lenticules behave differently from the 
axial ones. 

One solution to these difficulties is to project the images upon a diffusing 
screen, to fuse the elements together, then photograph the colored image 
that has been projected on the screen. This scheme is utilized by A. S. Howell 
(U.S.P. 1853683) and by H. E. Ives (U.S.P. 1985730 and 1985731). To re- 
move all traces of moire, the lenticulations on the master film are made 
to run at right angles to those on the copying material. This solution, while 
overcoming the flaw mentioned above, introduces a few new ones. Only a 
small fraction of the light projected upon the screen is picked up by the 
copying system, so that exposures are long. Also the diffusion that is intro- 
duced by the screen is sufficient to make the definition of the copied image 
distinctly inferior. 

The Eastman Kodak Company, very much interested in lenticular processes 
since they introduced it to the 16 mm field, offered several solutions to the 
problem of duplication. Dr. M. W. Seymour (U.S.P. 1708370 and 1912661; 
Eng. P. 382974) proposed to use curved gates both in the camera and in the 
printer. The curvature was predetermined so that when the film was sup- 
ported therein, the microscopic elements would project the images behind 
each lenticule at identically the same areas in the copying system, as in the 
taking. 

m 



250 



HISTORY OF COLOR PHOTOGRAPHY 



Another person to take advantage of the fact that the lenticules are minia- 
ture lenses, was P. Brosse (U.S.P. 1878670; Eng. P. 303356). This scheme 
is illustrated in Fig. 85. The master film is illuminated by means of diffuse 
light from a lamphouse B, containing lamps Z4, and L2, which illuminate a 
white reflecting curved surface S. The lenticular lens elements project this 



la.hykouse.% 




cctytre. cfj 
tncrror 

o 

Cof>yfi/m. 




FIG. 85 



light to a spherical mirror C, whose optical center is at 0. This is in the plane 
containing the two lenticular films A and D (master and copy), and is sym- 
metrical with respect to the exposed sections on each. After reflection from 
the mirrored surface of C, the rays coming from any point on A will recombine 
at the conjugate point on D. A special gate in front of films A and D pre- 
vents extraneous light from interfering. 

year 
•fbca>L 

of L2 




FIG. 86 

The optical methods to solve the problem of duplication all make use of a 
few simple principles. One of these is that if a lens L2, be placed so that its 
nodal plane of admission coincides with the rear focal plane of another lens Li, 
then the lens £2 will" project upon its own rear focal plane the image of the aper- 
ture of the first lens system. Thus the rays A y B,C, etc. (Fig. 86), all of which 
arise from the same point on the original subject, enter the lens system in 
the form of a beam.' Of these rays, A> B and H are intercepted fey the 'dia- 
phragm of the lens I*, the remaining rays proceeding through to the rear 



DUPLICATION OF LENTICULAR FILM 



251 



focal plane, where they converge to a point. If a lens £2 is placed at this point, 
the rays will be projected upon the rear focal plane of £2, in the exact image 
of the diaphragm of I*. This principle is, of course, the basic one which is 
utilized in the taking part of the process, but the lenses that are placed in 
the rear focal plane of the camera objective are the miniature lenses formed 
by the lenticulations. 

6otn /ettses 




plane, oflz 



FIG. 87 



Instead of placing lens system L2 so that its nodal plane of admission coin- 
cides with the rear focal plane of £*, it is possible to position it so that its front 
focal plane coincides with the rear focal plane of Li. The results, however, 
are the same, for upon the rear focal plane of the second system will be pro- 
jected the image of the aperture of the first system. This is made evident 
in Fig. 87. The rays A, B, C, etc., entering lens L\ in the form of a beam, 
are brought to a point focus in the focal plane of I*. But this is also the 
front focal plane of Z2, and as far as this lens is concerned, the rays originate 
at a point in its front focal plane. Therefore they are projected by the lens 
L2 as a beam of parallel rays in the exact image of the aperture of the lens L\, 
which determined their shape. 





A • , 


/ I 


PL 


— ■ — <- J h 


% J* c t 


£' ~- y T\ tf 




& < ^-^-^^ 4^A 5 


^ * ? 


/ ^ — jr 


p 5 


Ej. f 


ft c 


1 


s c 



FIG. 88 



When a lenticular film is illuminated from behind, depending upon the nature 
of the beam, the light that passes through any one lenticule will be either 
divergent or convergent. If the illuminating light is collimated, so that all 
the rays are parallel, then the beam is convergent, the point of convergence 
being in the front focal plane of the lenticular lens (Fig. 88). The rays of 



252 



HISTORY OF COLOR PHOTOGRAPHY 



collimated light A y B,C, etc., strike the lenticular film at right angles. There- 
fore they are not deviated until they reach the curved interface between len- 
ticular surface and air. They are converged to a point 0, which lies in the 
front focal plane of the lenticular element immediately facing the image points 
r, g, and b. If the point lies in the rear focal plane of another lens system, 
the rays will pass beyond in the form of a divergent beam, and form an 
image in the nodal plane of admission (lens aperture) of the new lens system, 
of the images r, g, and b. The rays will emerge from the lens in the form of a 
beam whose cross section, taken normally to the axis, will be an accurate 
image of the densities r, g, and b. 

It is also possible to use as a source of illumination, a cone of light whose 
apex is in the nodal point of the lenticular lens (Fig. 89). In this case the 




light emerges from the system as a divergent beam. This principle is utilized 
by O. E. Miller of the Eastman Kodak Company, for the projection printing 
of lenticular film (U.S.P. 2095826). The image of the light source S, is pro- 
jected by the lens system L\ y upon the master lenticular film which is placed 
behind a slit that is just sufficient to allow a single lenticule to be illuminated 
at any one time (Fig. 90). This is the lenticule A, situated on the optical axis 

dupt icalk, 
master 




Af/ttr 
source 



FIG. 90 



of the entire system. The nodal plane of lens A lies in the focal plane of lens 
£2, therefore, in accordance with the principle enunciated above, this lens 
will project in its own rear focal plane the image of the aperture of the lens A . 
But this image will correspond identically to the three densities that lie in 
the emulsion layer directly behind lens A, Thus at B there will be produced 
the duplicate of what lies in the rear focal plane of A, Since there will be an 



DUPLICATION OF LENTICULAR FILM 



253 



inversion of the image at B, the copy film could travel in the opposite direc- 
tion from the master. This will at the same time provide a slight relative 
omtion between the two which will destroy any moire pattern that may be 
inclined to be formed. 

A somewhat more complicated system is described by G. Chretien (Eng. P. 
286684). The objective has three separate positive optical systems, 0, C, 

C 




FIG. 91 

and 0', so arranged that the diaphragm D of the system 0, has for its image 
(beyond the intermediate system C), the anterior image D r of the lens system 
0'. Thus D and D f are conjugate (Fig. 91). Here again it is seen that C, 
being in the focal planes of system and 0', will project an image of the 
diaphragm of in the nodal plane of admission of the system 0'. This of 
course, coincides with the position D' y the diaphragm of the lens system 0'. 
If the diaphragm D is illuminated by a cone of light originating from a single 
horizontal or vertical lenticule, then the diaphragm D' registers accurately 
the intensity of the cone of light, and it will accurately image this intensity 
in its own rear focal plane in which the copy film is situated. 

A similar idea is disclosed by the I.G. in a series of patents (U.S.P. 1915418; 
Eng. P. 356701, 364559, 369262, 399984, .and 449954). An objective is used 




FIG. 92 



which is equal in function to the original taking lens. This means that the 
images of the set of filters, when viewed from the rear focal planes of both 
systems, will appear to be situated the same distance in front of these planes. 
One of the schemes is outlined in Fig. 92. A luminous plane E is situated 
in the focal plane of a lens system I*. Every point on E can be considered 
as a source of light. Consider the cone of rays originating at the point S. 
Since S is in the focal plane of Li, this lens will project the cone originating 



254 HISTORY OF COLOR PHOTOGRAPHY 

at S, in the form of a beam of light with parallel rays. Three of the rays 
are marked a h <h, and a 3 . These proceed to a lens system Li which is equal 
in function to the original taking lens. Since the rays #i, <h, a 3 enter lens system 
Li in the form of a beam, they will be brought to a point focus in the focal 
plane of Z2. At this point the master lenticular film is placed. 

The aperture of the lens system L2, is therefore illuminated by a whole 
series of beams each consisting of a pencil of parallel rays, and each making 
a different angle with the optical axis. These beams originate from the dif- 
ferent points on the luminous plane E. Since the lens Z2 is equal in function 
to the original taking lens, each one of these beams will be brought to a focus 
on a different lenticular element on the film N, and when viewed from these 
lenticulations each cone of light will have the same angularity that the corre- 
sponding beam had in the original taking system. Each beam will also cover, 
behind each lenticule in N, the same relative area that was covered in the 
original. The two systems differ in but one important respect, and that is 
that in the original system the intensities of the beams were determined by 
the reflection characteristics of the original. In the present case the intensities 
of the beams are all equal. Hence in the focal plane N, there is projected a 
uniform sheet of light, every portion of which makes the same angle with the 
optical axis that the corresponding portion made in the original. 

Now place at N, the master positive film. Let us consider the path of the 
cone of rays originating at the point S. After passage through L 1} these 
will emerge as a beam of rays that are parallel to each other, and to the optical 
axis. This beam is equivalent to a beam which originated at a point on the 
original object that lay on the optical axis of the original taking system. 
Upon passage through that system, one portion of the rays went through the 
red filter, another through the green, and the rest through the blue. In the 
copying setup, the equivalent portions are designated by the letters ai, <h, 
and a 3 , and for the sake of argument let us assume that the rays a h went 
through the red, the rays a 2 through the green, and the rays a 3 through the blue 
filter areas. The intensities of the rays after they passed through the filters 
were different, and depended upon the spectral composition of the beam of 
light of which they are the components. The three rays gave rise to three 
different densities behind the lenticule Si in the original. 

In the copying system, the rays a h 02, and a 3 are reconstructed in every 
particular except that the spectral compositions and the intensities of the three 
are identical. The angle that a x makes with the optical axis is identical with 
the angle made by the red rays after passage through the filter in the original 
taking system. Therefore the rays constituting a\ will come to a focus in 
the same relative area behind the lenticule S\ that the red rays did in the 
original. The same is true with the rays <h and a 3 . If the master lenticular 
film is placed at N, then the rays a h 02, and a 3 will become modulated upon 
passage through the lenticule £1 to a degree that corresponds to the original. 
Therefore the rays ai, (h r > and a 3 ', immediately after their passage through 



DUPLICATION OF LENTICULAR FILM 255 

Siy are no longer identical in intensity, but have intensities that correspond 
to those of the red, green, and blue portions of the corresponding original 
beam. When viewed from a point to the right of iV, the original scene is 
reconstructed in so far as light intensities are concerned. It becomes necessary 
merely to rephotograph this scene in order to make duplications. This is the 
function of the lens system Z, 3 and £4, in conjunction with the lenticular film P. 

Here again, advantage was taken of several optical principles in order to 
make the reproduction easier. Ordinarily, if an object A is to be copied one- 
to-one, it is placed in front of a lens system so that the distance between 
it and the nodal plane of admission is twice the focal length of the copying 
lens. At a distance twice the focal length away from the nodal plane of emis- 
sion, is placed the copy material. But it is extremely difficult to copy accurately 
silver images that are of the same order of magnitude as the disc of confusion 
of the lens and the resolving power of the copy material. It is much easier 
to copy a beam of light making a definite angle with the optical axis. This is 
the dodge adopted above. Consider the beam of light which is exemplified 
by the rays a h 02, a Sj at the moment it enters the nodal plane of admission 
of the lens L2. If it is desired to photograph the image of this plane on a 
one-to-one ratio, it becomes necessary to place a lens £3, where its nodal plane 
of admission is a distance twice its focal length from Fi, the nodal plane of 
emission of L2. This will project an image of the aperture Fi upon a plane F2, 
which is twice the focal length of the lens P 3 away from the nodal plane of the 
lens L$. The plane F 2 is the rear focal plane of £3 in the position occupied 
by Lz under the present circumstances. The intensity of the ray a{ as it 
enters the plane F 2 , corresponds to the intensity of that portion of the original 
beam which passed through the red filter. Hence at F% the beam is recon- 
structed identical in all respects except that of spectral composition; and this 
plays no role here, since the emulsion on the film P is sensitive only to the 
blue. If a lens Z4 is placed at F2, it will project upon a film P, in its own focal 
plane, the image of the aperture of the lens Z 3 . If the film at P is a lenticular 
film, the lenticulation, being in the rear focal plane of the lens L4, will image 
the aperture of £4, thus reconstructing the densities as they are at N. Since 
the rays at N and P are reversed with respect to each other, the film at P 
could move opposite in direction to that of N. 

A different scheme, but based upon the same principles, is outlined in 
English patent 295313. In Fig. 93, a condensing lens B registers the image 
of the light source Ay in the nodal plane of admission d of the compound 
lens D, which is equal in function to the original. The diaphragm d is movable. 
Placed immediately behind D is a lens E, whose characteristics are such that 
the focal length of the combination of lenses D and E is one-half that of D 
alone. Hence if the copy film F be placed in the rear focal plane of D, con- 
sidered individually, it is a distance twice the focal length of the system D 
plus E. If D is accurately focused on C, then the images at F and C will be 
identical in size. The size and dimensions of E must be such that no rays 



256 



HISTORY OF COLOR PHOTOGRAPHY 



through d are intercepted by E. As the individual rays originating at A pass 
through the film C, they become modulated so that the rays reaching F have 
the proper intensities. F and C being conjugate to each other, the lenticules 
at F will reconstruct the images behind the corresponding lenticules at C. 
Hence duplication is achieved. Abbe's experiments have shown that if d 




FIG. 93 

allows only the central image of the light source to pass through, the image 
of the network at C cannot be formed at jF. Hence no moire patterns can 
arise. This scheme is also suitable for use in duplicating screen plates. 

A very popular type of printing objective utilizes a symmetrical lens. This 
can be formed by coupling two identical lens systems. Because of the com- 
plete reversibility of light paths, one lens will completely reverse the other. 
The aperture of the coupled system would be exactly in the centre. A system 




such as this has been disclosed in English patent 274848, issued to "Societe 
Civil pour l'fitude de la Cinematographic en Couleur." A master lenticular 
film A, is illuminated by means of light C (Fig. 94). The front portion of the 
symmetrical lens system D starts to project this image, but the rear portion 
reverses the procedure and reconstructs the rays at B, the copy film. The 
films B and A are conjugate to each other, so that the miniature lenses at B 
will reconstruct the apertures of the miniature lenses at A. The aperture E, 
of the lens system D, corresponds to the position, with respect to the front 
portion of D, that would be occupied by the filters were D allowed to pro- 
ject the image in the normal manner. Hence E also corresponds to the position 



DUPLICATION OF LENTICULAR FILM 257 

of the filters, as far as the rear portion of the system D is concerned. The 
lens D is characterized by two properties: 

1. The aperture E, of the system D, must afford passage to a pencil of light 
rays whose angle must be the same as that of the corresponding beam 
in the original taking system. 

2. None of the rays received by the lens will be intercepted before reaching 
the copying film, B. 

F is a lamphouse that fogs the copying film B, so that color dilution, due to 
the blocking of light by the interstices between the lenticular elements at A, 
will be overcome. 

Dr. M. W. Seymour, of the Eastman Kodak Company, likewise used a 
symmetrical lens system (U.S.P. 1976300). In the plane of the common 






diaphragm to the two portions of the symmetrical objective, is placed a special 
diaphragm, thus enabling him to place the lenticulations of the copy at an 
angle to those of the master. This reduced moire. In lenticular processes, 
it is essential that the filter bands be parallel to the lenticulations. In Fig. 95, 
let A be the filter bands parallel to the direction of the lenticules in the master, 
and B the filter bands parallel to the direction of the copy. Then C will be 
the shape of the diaphragm used. In the diagrams A and B, the shaded 
portions indicate different colors. But in diagram C the shaded areas repre- 
sent opaque portions, while the clear areas represent transparent portions. 
J. L. Vidal (Eng. P. 349276); A. Dervieux (Eng. P. 451283); K - Rantsch, 
of the Siemens and Halske Company (U.S.P. 2009816); the Societe franchise 
de Cinematographic (Eng. P. 294579); the Keller-Dorian Co. (Eng. P. 284995); 
and F. E. Tuttle, of the Eastman Kodak Company (U.S.P. 2039691; Eng. 
P. 411407 and 446752) also disclosed similar ideas. 

The Tuttle disclosures deserve some special attention because a principle 
in optics is disclosed that is used considerably in other schemes. The printing 
is done in three stages, one color at a time. A different light source is used 
for each color. These are marked numbers 17, 18, and 19 in Fig. 96. Consider 
first the case where the printing is done by means of light source 18. This 
lies on the optical axis of the entire system. In front of the three light sources, 
is placed a diaphragm shutter (20) which is coupled with another shutter 



2^8 HISTORY OF COLOR PHOTOGRAPHY 

(21) placed on the same shaft, and which rotates between the two symmetrical 
elements of the copying objective (16). The plane containing the light sources 
17, 18, and 19 lies in the front focal plane of the collimating lens (23), behind 
which is placed the master lenticular film (10). Immediately behind the 
film is a lens (24), which directs the rays to the symmetrical copying objec- 
tive (16). The master film, lying between the two lenses (23 and 24), acts 
like a diaphragm for the combination, modulating the intensities of the in- 
dividual rays. The copying objective (16), projects an image of these rays 
upon the copy film (13), through a compensating lens (25). 




FIG. 96 



Since the light source (18) lies on the front focal plane of the lens (23), 
and on its optical axis, the cone of rays emanating from this point will be con- 
verted by the lens (23) into a beam that is parallel to the optical axis. Therefore 
every lenticular element in 10 will be illuminated by a pencil of rays that are 
parallel to the axis of the lenticular lens. Each pencil of rays will be brought 
to a focus on only one image point behind the lenticular element, this being 
the image of one color density. In a similar manner, the cone of rays starting 
at 17 or 19 will also give rise to beams with parallel rays, but these make 
different angles with the optical axis. Therefore these will illuminate the 
lenticular elements with pencils of rays that make different angles with the 
optical axes of the miniature lenses, and they will be brought to foci. at different 
points behind the lenticules. The positions of the points 17, 18, and 19 are 
such that the pencils of rays will assume the same angularity with respect 
to the optical axis that was originally assumed by the rays after they passed 
through the red, green, and blue filters in the original. Each pencil will be 
brought to a focus at a point behind the lenticular element, corresponding to 
a single color element. 

In Fig. 96 only the axes of the pencils passing through each lenticular 
element have been indicated, hence the reason for the presence of a diaphragm 
shutter (21), positioned between the elements of the copying objective (16), 



DUPLICATION OF LENTICULAR FILM 



259 



is not clear. Let us consider the case of the entire pencil of rays, taking into 
consideration also, the optical properties of the lenticules. This is indicated 
in Fig. 97. The cone of rays originating at 17 becomes converted into a beam 
with parallel rays. The lenticules in 10 become illuminated by the parallel 
rays, hence they will bring the rays to point foci in the image plane of 10. 
There the pencils become modulated by the image due to a single color ele- 
ment. After passage through this point, the rays diverge again into a broad 
pencil. The lens (24) directs them into the copying objective. Here they 
cover much more than the area that would normally correspond to a single 




FIG. 97 



beam, hence the diaphragm (21), coupled with 20, eliminates all those rays 
which extend beyond the proper confines. Therefore the lenticulations at 
13 become illuminated by cones of light whose axes make the same angles with 
the axes of the lenticules, as the rays through the red filter made with the 
lenticulations in the original exposure. 

This principle of illuminating the master lenticular film by three different 
light sources, so arranged that the optical system converts the rays from these 
sources into three pencils bearing the same angularity to the lenticular axes 
that the filter areas bore to them in the original taking, is disclosed in a num- 
ber of patents issued to J. Eggert and G. Heymer (U.S.P. 2013178; Eng. P. 
367414 and 399977). Three gas-discharge lamps, or one lamp with three fila- 
ments, are used, so positioned in front of a lens system that their images 
appear to occupy the positions occupied by the filters during the original 
exposures. 

The pencils of rays after passing through a plane that lies in the focus of a 
lens system, will diverge again into a broad beam. Wherever complicated 
optical schemes are used, such as the ones disclosed by Dr. Tuttle, the breadth 
of the beams may be such that overlapping of images occurs. It is to be 
recalled that he introduced a. special diaphragm to confine the limits of these 
beams. Dr. Tuttle was preceded in- this by P.-Fournier (Eng. P. 236204) 



260 



HISTORY OF COLOR PHOTOGRAPHY 



who utilized only that portion of each beam that corresponded to the central 
portion of the filter areas. G. Heymer (U.S.P. 1915418; Eng. P. 400057) 
also adopted a similar procedure. 

The use of a separate pencil of rays to print each color element behind each 
lenticule, is disclosed by K. Rantsch in a whole series of patents, the earlier 
of which were issued to the German engineering concern, The Siemens and 
Halske Company (U.S.P. 2022978 and 2036499), and the later ones to Opti- 
color A. G. (U.S.P. 2061088, 2070179, 2071764). All the corresponding English 
patents were issued to Opticolor. These bear the serial numbers 421084, 
422983, 422991, 430467, 430503, 435&59, 436465, 436466, 436608, 4383S6, 
438748, 440273, 440809, 441028, 441709, 443266, and 508219. In general, 
these disclose a system where a light source is situated in front of a system 
of mirrors which directs three pencils of light rays, at different angles, into 
three objectives which in turn direct the rays to the same point in the common 
focal plane of the lenses. The master lenticular film is placed between the 
light source and the mirrors, so that for all intents and purposes this can be 
considered as the light source. Every lenticular element on the master film 
will direct a pencil of rays to each of the mirrors and objectives. The mirrors 
are so positioned that each pencil represents the rays proceeding from one 
of the three color densities behind each lenticule. 

A rather simple scheme is outlined in Fig. 98. If the master lenticular film, 
M, is illuminated with diffuse light, each of the densities behind each lenticule 




FIG. 98 



becomes the origin of a point source of light. Due to the optical properties 
of the lenticules, the ray from the points a,\ and 02 behind a single lenticule, 
converge to a common point from which they then diverge. Therefore it is 
seen that the light from the three-color densities gives rise to three beams of 
pencils of light, a h <h, clz (only two of these are indicated in the diagram). 
By a proper choice of the mirror reflectors Ri and R 2 , these widely divergent 



DUPLICATION OF LENTICULAR FILM 261 

rays are reunited upon a single lenticular element in the copy film C. Since 
the elementary lenses see only three pencils of rays converging upon them, 
each of the pencils will be imaged as a distinct point behind each lenticule. 

A somewhat more complicated arrangement which accomplished the same 
result is depicted in Fig. 99. Here again the light originating at the three 
image densities behind each lenticule, gives rise to three pencils of rays which 
intersect each other in the focal plane of the lenticules of the master film M. 
The widely divergent pencils (two of which are indicated in the diagram) 
a\ and 02, are then reunited by a system of lenses L\ and L% and reflectors Ri, 




FIG. 99 

R2, Rz, and R*, so that they reform at a single lenticule on the copy film C. 
Since this lenticule sees the rays from widely differing angles, it will image 
them as separate densities. In United States patent 2062146, K. Rantsch 
discloses some other very general processes. 

The Keller-Dorian Company (Eng. P. 247168) disclosed a projection printer 
which used an objective whose pupil of emergence was at infinity. Therefore 
the images of the three filters in the lens aperture are independent of the 
focal lengths of the objective, since the rays passing through them emerge 
as three beams, each containing rays parallel to each other, each beam making 
a different angle with the optical axes of the lenticular elements. F. Strecker 
(U.S.P. 2040280) would place filters in the copying objective to correspond 
to the filter areas in the original objective. But these filters would pass the 
ultraviolet, violet, and blue-green regions respectively. Therefore ordinary 
colorblind emulsions could be used for copying, since the sensitivities of such 
materials extend slightly into the green, just beyond the blue-green. In order 
to prevent the light from a given point in the master from reaching two lenticu- 
lar elements in the copy, F. Fischer (U.S.P. 198447 1) placed a grid or other 
refracting surface, in front of the silver image. Anne Henri Jacques de Lassus 
Saint Genies disclosed several projection printing schemes in a series of patents 
(Eng. P. 457656, 460533, 462996, and 464723). These appear to be modifica- 
tions of schemes already discussed. Other schemes were disclosed by A. Blon- 
del (Eng. P. 462655); E. M. Sandoz, (Eng. P. 495348); Kapella, Ltd. 
(Eng. P. 373938); E. Leitz (U.S.P. 2055237); and C. Nordmann (Eng. P. 



262 



HISTORY OF COLOR PHOTOGRAPHY 



364627 and 410609). The last two are important in that they discuss funda- 
mental principles from the point of view of general optics, and they lay down 
the conditions that must be fulfilled when the image sizes or the optical 
characteristics of original and duplicating systems differ from each other. 

The usual conditions met during the taking of photographs with lenticular 
film are depicted in Fig. 100. A pencil of rays a h 02, a S} enters the lens. In 
the aperture of the objective, is positioned a filter F containing three filter 
zones r, g, and b, colored red, green, and blue respectively. After passage 
through the lens, the three colored pencils ai, 02', #3', are recombined upon 




FIG. 100 



the front surface of the lenticular film M> and give rise to three images ti, {2, 
and iz, in the emulsion layer of M. To reconstruct the image in its original 
state, the paths of the light rays must be identically reversed. 

In order for this to happen, it becomes necessary and sufficient that the 
light which illuminates the film from behind be broken up into three pencils 
by each set of image densities behind each lenticule. The ray in each pencil 
proceeding from the center of the image to the point on the lenticule which 
represents the nodal point of emission for the miniature lens, can be con- 
sidered as the elementary axis of the pencil. Therefore from each lenticular 
element, there proceed three axes. This is illustrated in exaggerated form 
in Fig. 101. The relationship of the image dots i h i^ H, nh, nh, m%, to the 
lenticular elements A and B is determined by the shape and number of lenti- 
cules in the master film M, and by the optical properties of the lens through 
which the initial exposure has been made. It is the focal length of the lens 
and the dimensions of the filter bands occupying its aperture, which deter- 
mine the angularity which the rays b h b 2 , h, a h a%, a 3 (Fig. 102), make with 
the optical axis of the entire system. But one thing is true regardless of what 



DUPLICATION OF LENTICULAR FILM 



263 



the lens characteristics are. The rays a x and b h which pass through the same 
section r of the filter band F, in the aperture of the lens system L, intersect 
each other in the exact center of the filter zone r. The same is true of the 
rays 02 and J 2 , and the rays a 3 and fa. We can therefore make this very general 




/^"IVt 



FIG. 101 



statement: When a lenticular film is illuminated from behind, the elementary 
beams proceeding from corresponding points behind each lenticular element, 
will, in the absence of an optical system, intersect each other at points which 
correspond to the center of the filter zones present during the original pho- 
tography. The distance x, of these points of intersection, from the lenticular 
surface of the film is therefore a definite characteristic of the master film, 




FIG. 102 

and is fixed by the optical properties of the lens system in the original camera 
exposure. If it is desired to project a lenticular film in natural color, it is 
necessary and sufficient that the filters be placed in the plane of these points 
of intersection, and the centers of the filter zones must coincide with these 
points. If there is interposed an optical system in the path of these rays before 
the points of intersection, and if the nodal planes of admission and emission 
are separated by a distance equal to /, then the points of intersection of 



264 HISTORY OF COLOR PHOTOGRAPHY 

the elementary axes are moved forward a distance equal to Z, so that the 
new position of the filter zones is at a distance x + I from the lenticules. 

Since the optical properties of the projection system are always different 
from those of the original taking system, it is not always possible to have 
the plane of the points of intersection clear so that a filter system can be 
installed there. To overcome this difficulty, Keller-Dorian (Eng. P. 247168, 
cf. above) proposed to make these axes parallel to each other, by placing a 
special optical system between the color filters and the film not only during 
the copying, but also during the projection. In this case the exit pupils will 
intersect each other at infinity. This makes the system independent of the 
focal length of the objectives used. In order to project this film, it becomes 
necessary to interpose between the film and the aperture of the projection 
lens which contains the filters, such an optical system that the elementary 
axes will intersect in the centers of the filter zones. A system of this type 
is known as a collimation system. 

The Nordmann disclosures represent a different attack upon the same 
problem, and one which does not utilize the expensive added optical equip- 
ment. It is necessary to know the position of the filters in the projection 
system. Suppose this be a distance of y + 1 from the lenticular surface, 
where I is the distance between the nodal planes of admission and emission 
of the projection objective. In making copies, it is necessary to use an optical 
system which will cause the elementary axes to intersect at a point y in front 
of the lenticular surface. If the master lenticular film is such that these axes 
intersect each other at a point X, which is a distance x in front of it, then 
a definite relationship must hold between X and F, the point of intersection 
of the corresponding beams in the copy material. It is known that X is a 
distance x from the lenticular surface of the original, and Y is a distance y 
from the front surface of the copy film or a distance y + l if an optical system 
lies between Y and the copy film, I being the separation between the nodal 
planes. The relationship is that the points X and F are conjugate to each 
other. This means that the rays that go through X, must coincide with rays 
that would ordinarily go through F. 

This will become clearer as we analyze the problem in greater detail (Fig. 
103). It is desired to make copy and master identical in size. Hence if the 
focal length of the lens system used in copying is/, the master film C is placed 
a distance equal to 2/ in front of the nodal plane of admission Pi, of the system. 
The copy film /, is placed a distance 2/ behind the nodal plane of emission 
P 2 of the system. The planes P x and P 2 are separated a distance I from each 
other. The axes of the elementary beams from the lenticules A and B, inter- 
sect at three points which represent the centers of the filter zones. Consider 
the rays that would normally go through the central zone. These rays will 
intersect at a point X, which is a distance x in front of the plane containing 
A and B. The ray a, from the lenticule A, after going through the point X 
continues until it intersects the nodal plane Pi at the point S. The ray b 



DUPLICATION OF LENTICULAR FILM 



265 



from the lenticule B, intersects the nodal plane P± at R. The two rays then 
translate themselves parallel to the optical axis 00', until they reach the 
nodal plane of emission P 2 . This they do at the points R f and S'. From these 
points the rays b' and a', continue until they reach the rear focal plane of the 
lens system. To determine the exact points A' and B' where these rays inter- 
sect the rear focal plane, we draw from the points A and B the rays c and d, 
which go directly to the nodal point of admission M. From here the rays 
translate themselves to the nodal point of emission N, then continue in the 




original directions until they intersect the rear focal plane. Thus it is seen 
that the rays from the points A and B y after going through the point X, will 
enter the lens system and give rise to the points A 1 and B r upon the copy 
lenticular film. 

Now consider the copy lenticular film 7. If this is illuminated from behind, 
the central elementary axes will intersect at a point Y! which is a distance 
y in front of the lenticulations. This is in the absence of a lens system be- 
tween the two films. In the presence of a lens system, the distance of the 
point of intersection will be moved forward a distance /, which is the separa- 
tion between the nodal planes of admission and emission. This is the point 
Y in Fig. 103. Let us illuminate the two films C and I, both from the side 
carrying the images. Let us center our attention upon the central pencils 
of rays through each lenticule. The axes of these pencils will intersect at 
the point X or the point Y, depending upon whether they originate at C or 
I. We say that X and Y are conjugate to each other if the following conditions 
hold: A' and B' are the images of the points A and B. The axes of the central 
pencils of rays from the lenticular elements A and B, are the rays a and b. 
The axes of the corresponding pencils from A 1 and B 1 are a! and b'. The rays 
a and b intersect at the point X since no optical system lies between X and C. 
The rays a! and b', in the absence of such a system, would intersect at the 
point F', but since there is a lens between I and the point of intersection, 



266 HISTORY OF COLOR PHOTOGRAPHY 

this moves forward to the point F. If the rays a and b, after passing through 
the lens system, coincide with the rays o! and V y then the points X and Y 
are conjugate to each other. These conditions determine x and y. Since Y 
is the virtual image of the point X, the relation 

P^q f 

must hold, where p is the distance of the point X from the nodal plane of ad- 
mission, q the distance of the point Y from the nodal plane of emission, and 
/ is the focal length of the lens. The distance p is XM, and this is equal to 
2/ - x. The distance q is YN or YL - NL. But NL is 2/ so that q = y - 2/. 
But g is positive if it is measured to the right of N, and negative if it lies to 
the left, hence q is negative. Substituting these values in the equation above, 
we have: 

T 1 1 

H 7 7v = -j. or 



2f-x - ( y - 2/) / 



2 / ~ J / * - 2/ 

In this equation the value of x is known, since it is a characteristic of the 
master film. It is equal to the distance between the lenticulation and the 
plane of the filters in the camera during the original exposure. The value of 
y is also known, since this is the distance between the plane of the filters and 
the lenticulations in the projection system. Therefore there is but one un- 
known, /, the focal length of the lens that is used for the copying system. If 
during the original exposure the object was at the infinity distance, and if the 
screen projection is arranged so that a life-size enlargement is obtained, and 
if the filters are placed in the front nodal planes of the systems, then x is the 
focal length of the original lens system in the camera and y is the focal length 
of the projection system. Thus these three optical constants are related 
to each other in a very simple manner. 

In the second of his patents (Eng. P. 410609) Mr. Nordmann takes up 
the more general problem of making duplicates a different size from the 
original. Thus it may be desirable to make the initial exposures upon 35 mm 
film, and the duplicates upon 16 mm material, or vice versa. With the recent 
introduction of 8 mm film as a standard material, this field is greatly enlarged. 
The relationship that holds now is the following: 

1 . * I j 

s p 

where p is the distance from the master lenticular film C (Fig. 104) to the 
nodal plane of admission, p' the distance of the copy lenticular film I from 



DUPLICATION OF LENTICULAR FILM 



267 



£*_/>'=/££_* 




FIG. 104 



the nodal plane of emission, / the focal length of the lens system, and g the 
degree of enlargement. Let us suppose g = J, then 

p' 



- = — or 



P ' = ip 



The master film is twice the distance from the nodal plane of admission that 
the copy film is from the nodal plane of emission. Substituting this value 
for p f into the lens equation, there results: 

P + fp = 7 p~7 p 3/ 

Therefore the master film is placed a distance equal to three times the focal 
length *in front of the nodal plane of admission, while the copy film is placed 
one and one-half times the focal length behind the nodal plane of emission. 
For a 2-inch lens, these are six and three inches respectively. Between the 
point of intersection X of the elementary axes of the central pencils of rays 
from the lenticular elements A y B, etc., of the master C and the copy film 7, 
is placed an optical system whose focal length is /, and whose nodal planes 
P1P1 and P2P2 are a distance I apart. As before, it is desired to use a pro- 
jection system in which the filters are placed a distance y + 1 in front of the 
lenticular surface. This fixes completely the focal length / of the copying 
system, since this must be such that the axes of the central pencils of rays 



268 HISTORY OF COLOR PHOTOGRAPHY 

from each lenticule A', B', etc., must intersect each other at a point Y> which 
is a distance y from the lenticular elements, when no optical system is inter- 
posed between Y and the film. In the presence of such a system the point Y 
is moved to Y\ a distance equal to the separation between the nodal planes. 

The points X and Y f are conjugate to each other. Therefore the standard 
lens equation is applicable to these points. The distance X from the nodal 
plane of admission is XN, which is equal to RNi — R or p — x. The distance 
Y' from the nodal plane of emission is Y'N* which is equal to Y f S — N2S. 
But in the distance Y'S is included the distance between the two nodal planes, 
I. From the point of view of optics, this distance is completely neutral, 
hence must be subtracted from the total. Y'S = y + l, hence the optical 
distance is y. The distance N2S is p'. Therefore the equation may be written: 

+ 



p-x p'-y f 

If the point X lies on the other side of iVi, x will be greater than p, and the 
relationship will then take the form: 

1.1 1 



x-p y-p' f 

Since s and y are known constants, and p 7 p' and / are related to each other 
(p = 3/, p' = i|/, for a reduction to \ size), only the value of / is unknown, 
and determinable from the equation : 

1.1 1 

= -. or 



*-3/ y-if f 

-i— + 2 ,i 

* - 3/ *y - 3/ / 

The patents by Nordmann are very general in scope. They disclose the 
most general conditions that have to be satisfied in order to make duplicate 
copies from a master lenticular film. The original taking system may differ 
from the projection system not only in the focal length of the objective used, 
but also in the effective aperture of the lens, the width of the filter zones, etc. 
The I.G. in several disclosures (Eng. P. 434206 and 444191) outlines systems 
which compensate for these and other characteristics of the camera and 
projection systems. The Keller-Dorian Company (Eng. P. 297773) would 
solve most of these problems by first reproducing the images on non-lenticular 
film, then copying the reproductions upon lenticular film through an appropriate 
optical system, to yield the desired properties. In copying film of this type, 
it was found that the interstices between the lenticules in the master did not 
transmit light. This leaves the corresponding sections in the copy blank, 
leading to color dilution. The Keller-Dorian Company remedied this by 
giving the film a supplementary exposure through the lenticular elements 
using a special annular diaphragm (Fig. 105), having the central portion 



DUPLICATION OF LENTICULAR FILM 269 

completely blocked out. This portion corresponds to the space directly be- 
hind the lenticule. The clear portion would affect only the regions beyond 
the central zone (Eng. P. 2451 18). 

It is also possible to print the lenticular film by contact. Here the main 
problem is to avoid the formation of moire. This will take place whenever the 
pattern of the master fails to register completely with the pattern of the copy 
material. The effects of moire are identical in this case with those in the 
case of screen plates. It may be recalled that a very successful solution was 
to have the emulsion of the master separated from the emulsion of the copy 
film by the thickness of one film base. This has its counterpart in the schemes 




FIG. 105 

disclosed by the LG. where a glass plate with parallel sides (Eng. P. 31 7051), 
or a clear lenticular film (Eng. P. 435813) is placed between the master and 
copy, the entire group forming a sandwich. 

The printing of the copy upon a film whose lenticulations are at a definite 
angle to those of the master, has been proposed quite often by a number of 
people, among whom are H. Ami (U.S.P. 1876442); Eggert and Heymer 
(U.S.P. 2072396); Capstan" (U.S.P. 2058409); J.T,. Vidal (U.S.P. 1935422); 
the Keller-Dorian Company (Eng. P. 21 1486 and 391 no); the Societe 
franchise de Cinematographie (Eng. P. 31 7051 and 329899); and the Kislyn 
Corp. (Eng. P. 371546). In some of the disclosures the lenticulations were at 
right angles to each other, in others the angle could be as low as one or two 
degrees. 

The projection printing of lenticular film was also complicated by the 
problem of moire. The LG. proposed to place a piece of clear lenticular 
film in the path of the rays (Eng. P. 36741 1). This had also been proposed by 
O. E. Miller of the Eastman Kodak Company (U.S.P. 2008989). The Keller- 
Dorian Company disclosed many schemes to solve this problem. Direct 
vision prisms (U.S.P. 1982 187; Eng. P. 378249 and 391908), parallel plate 
glass at an angle to the optical axis (U.S.P. 1884996 and 1904671; Eng. P. 
288290 and 378245), prisms which deflected the images (U.S.P. 1884995; 
Eng. P. 265069 and 304643), doubly refracting elements (Eng. P. 374993) 
and diffraction gratings (Eng. P. 303357) were placed in the path of the 
light rays, to cause a sufficient deflection of the beam to remove the moire. 



270 HISTORY OF COLOR PHOTOGRAPHY 

The Societe frangaise de Cinematographic (Eng. P. 310320) and the Kislyn 
Corp. (Eng. P. 377717) proposed to reflect the exposing light off a rocking mir- 
ror. In another disclosure this company suggested that the copying objective 
could be vibrated (Eng. P. 309540). The Keller-Dorian Company used the 
inefficiencies of the optical system to eliminate the moire (Eng. P. 317060) 
operating under a circle of confusion that was beyond the structure of a lenticu- 
lar element. The I.G. (Eng. P. 434205) used a cylindrical lens in the copying 
system to effect a spread of image. Other proposals were made by Rantsch 
(U.S.P. 2011263) and S. B. Colgate (U.S.P. 2026376). 

In printing lenticular film by contact, the lenticules at the edges of the 
image receive light at a different angle from those in the center. Unless this 
is neutralized, the effect is to give the image a dominant line. C. Nordmann 
studied this problem and suggested a number of solutions (Eng. P. 360524, 
362490, 363387, 363409 and 363447). The I.G. also disclosed various schemes 
by which lenticular film could be printed by contact (Eng. P. 392987, 417020, 
417096, 440567 and 447317). 

The many intricate problems that beset the technician when he attempted 
to print duplicates upon lenticular film naturally forced his attention to the 
possibility of making color separations from a lenticular master film. The 
separations could then be used to prepare any type of color positive that was 
available to him. There were many advantages to this idea in the days before 
monopacks and condensed tripacks were disclosed. It made the camera man 
independent of the whims of balky and temperamental one-shot cameras and 
one-shot camera owners, and it was not necessary to carry three times the 
quantity of film. But best of all it enabled him to use the short-focal-length 
lenses to which he was accustomed in black-and-white work, thus making it 
unnecessary to shift his lighting arrangements. No registry problems re- 
mained to hound him. In short, all the headaches of color became removed 
to the precision laboratory where they belonged. 

A very obvious method of accomplishing the separation was to project a 
colored image upon a screen and copy the projected picture through the 
separation filters. This was disclosed by A. Blondel (Eng. P. 462794). Any 
of the proposals outlined above for the projection printing of lenticular film 
could be used also if special diaphragms were placed in the paths to block out 
the rays that would correspond to two primaries. Thus but one color is 
printed at a time. Similar proposals were disclosed by H. T. Kalmus and J. A. 
Ball (Eng. P. 343369); H. E. Ives (U.S.P. 1970936); and by the I.G. (U.S.P. 
1874529; Eng. P. 353121, 402902, 440187). 

A rather novel system, based upon the principle of masking, is outlined by 
the I.G. in a series of patents (U.S.P. 2064058; Eng. P. 412021, 428875, 
435994 and 440025). When a lenticular film is illuminated by means of a 
uniformly lighted area equal in size to one of the filter bands, and arranged 
at a substantial distance from the plane of the lenticules, there is obtained 
a line running parallel to the lenticules. The line will cover an area that is 



DUPLICATION OF LENTICULAR FILM 271 

one-third of the space behind the lenticular element, and the exact portion 
that the line covers depends upon the angle which the axis of the beam makes 
with the elementary axes of the miniature lenses. Since the source of illumina- 
tion is substantially at infinity, the light reaching the film consists of parallel 
rays. Three such films are made, one each with the angle to correspond to 
that of a single primary color. If these are printed by contact upon a smooth- 
based film, registry being maintained by use of the perforations on the side 
of the film, there will be obtained a line which corresponds to the width of two 
color densities. The contact prints are called stencils. The master positive 
or negative is now printed by contact upon a smooth-based film with a stencil 
lying between the master and the copy film. This will completely block out 
the densities corresponding to two color primaries, leaving the third one to 
come through to print upon the copy film. Since there will be the thickness 
of a base between the master and the copy, no trace of moire will be present 
if the source of illumination be properly chosen. This was discussed in some 
detail in the chapter dealing with screen plates. 

In order to avoid the lateral inversion that takes place when lenticular 
film is printed, G. Heymer (U.S.P. 1874529; Eng. P. 353121) would print 
the film with the emulsion sides in contact. If the printing is done by projec- 
tion the original objective can be used, and the zones occupied by two of the 
filters are to be blocked out. This is practically identical to the schemes pro- 
posed by Ives and Kalmus and Ball. P. A. Richard (U.S.P. 1750358) also 
disclosed a similar scheme. 

The reverse procedure, printing color separations on to lenticular positive 
film, has also been disclosed. Several schemes were proposed by V. Hudeley 
and J. Lagrave (U.S.P. 2030795; Eng. P. 408109, 421582 and 430585). The 
negatives were made in a special triple-lens camera, which gave three images 
in a single frame. This created the problem of enlarging and properly register- 
ing each of the three part-images in a single frame of lenticular film. The 
I.G. Company (Eng. P. 392987) would print the separations on to the lenticu- 
lar positive by contact. The color record is placed in contact with the lenticu- 
lar side of the copy film, and the beam used for exposure is given the proper 
angle with respect to the elementary lenticular axes so that the image will 
be formed in the proper zones behind each lenticule. In a subsequent patent, 
a stencil such as is disclosed in English patent 41 2021 is placed between the 
color record and the lenticulated side of the copy film. Here again the angu- 
larity of the incident light is controlled to deposit the image in the proper zone. 

Dr. Bela Gaspar (Eng. P. 343369, 406013 and 409270), and the I.G. (Eng. P. 
375338, 451175 and 460653) proposed to print lenticular master positives 
upon other types of monopack film, and process them to color transparencies 
by means of dye-destruction or dye-coupling methods. 

In line with the use of lenticular film as a negative process, the I.G. has 
disclosed its use as the front element of a bipack (U.S.P. 2093655; Eng. P. 
395124). The film is coated with an orthochromatic emulsion. In the lens 



272 HISTORY OF COLOR PHOTOGRAPHY 

aperture there is placed a two-banded filter with a yellow and a magenta zone. 
Both of these transmit the red rays. Since the front emulsion is not sensitive 
to the red, these will not register here, but upon the rear red-sensitized film. 
A continuous-tone record of the red densities will be obtained. The front 
lenticular film will register but two densities behind each lenticular element, 
since but two filter zones are present in the lens aperture. One zone transmits 
blue and red light, therefore only blue will be registered by this beam. The 
other zone transmits red and green, hence only green will be recorded by it 
upon an orthochromatic emulsion. The subsequent separation of the two 
color densities is a much simpler job than the separation of three densities. 



CHAPTER 19 
MASKING 



R 



I discussion of the negative processes would be complete without men- 
tion of the subject of color correction by masking. This technique was first 
introduced by Dr. E. Albert in the eighteen-nineties (Ger. P. 101379 and 
1 16538) to improve the qualities of the black printer in four-color reproduc- 
tion. The use of a key plate to give the final print more contrast and better 
blacks, had been firmly established about ten to fifteen years before this, and 
the practice introduced several problems that did not exist previously. One 
of these problems was that all the bright colors became degraded because 
black was printed into them by the key plate. Dr. Albert proposed to eliminate 
this trouble by making a weak positive from the black printer negative and 
placing it in registry with the three-color negatives. The combination was 
used to print the positives. This had the effect of removing black from each 
of the color-printing plates. The general procedure of combining a positive 
of one separation with a negative of another is termed masking. 

Another method for producing the same result was disclosed in the Process 
Photogram (Vol. 8 (1901), p. 114). Here it was proposed to combine each 
separation negative with positives made from the other two negatives. This 
was later patented by L. 0. van Straatten (Eng. P. 353 151). Schelter and 
Giesecke (Ger. P. 152799) patented the use of secondary masks; that is, 
masks made from negatives that were exposed behind magenta, cyan, and 
yellow filters. It is therefore seen that the van Stratten technique is but a 
modification of this one. In all of the disclosures made up to this point, the 
main objective was to eliminate black from the color printer. Thus the com- 
bination of a red negative with a minus-red positive will increase the densities 
in the masked negative at those points where green and blue densities are 
being deposited by the other two plates. Therefore less red densities (less 
cyan color) will be deposited here. Black, where it is required, will be de- 
posited by the black plate, so that no loss of density or definition will take 
place. The net effect will be to brighten up all the colors, thus increasing 
brilliance and saturation. 

It was O. Pfenninger and E. C. Townsend (Eng. P. 26608/10) who first 
proposed masking as a corrective for the poor transmissions of the magenta 
printing ink. We have met Mr. Pfenninger previously in our discussions 
concerning one-shot cameras. In order to compensate for the poor blue trans- 
mission of the magenta ink, they combined a weak positive of the green-filter 

273 



274 HISTORY OF COLOR PHOTOGRAPHY 

(magenta printer) negative, with the blue-filter negative to form a masked 
yellow printer. This was later patented by J. A. H. Hatt (U.S.P. 1349956), 
but in a different form. The green-filter negative was placed in a special camera 
which allowed registry to be carefully maintained. The red-filter negative 
was then projected upon this in exact registry. The position of the green-filter 
negative was carefully marked, and the plate then removed and coated with 
a collodion emulsion. It was finally replaced in the camera, and a red-filter 
positive printed upon it. By this means there was formed a green-filter 
separation that had a weak positive of the red-filter separation combined 
with it. The extension to three colors was proposed by J. Keenen (U.S.P. 
1866556). Mr. Keenen also pointed out that the masks could be made upon 
separate plates that could then be bound in exact registry with the different 
negatives. F, H. Hausleiter (Eng. P. 360441) also proposed a like technique, 
as did Wilkinson (U.S.P. 2004144); R. Mackay (U.S.P. 2060816); J. R. 
Elsworth (Eng. P. 440086); and Feeny (Eng. P. 444229). Other complicated 
arrangements utilizing special types of plateholders and cameras that allowed 
the removal and replacement of plates in exact positions, were disclosed by 
E. H. Gamble (Eng. P. 6768/12); E. A. Raschke (U.S.P. 1373020); J. A. H. 
Hatt (U.S.P. 1518426 and 1569124); S. J. van Straatten (Eng. P. 427234); 
L. Nerot (Fr. P. 780364); and Boettger and Kronschnabl (Ger. P. 592003 
and 601484; Eng. P. 402082; U.S.P. 2020688). A complete disclosure as 
to the working details of a masking process is contained in the Eastman 
Kodak booklet "The Modern Masking Method of Correct Color Reproduc- 
tion." Here is disclosed one novel feature. The black printer is prepared by 
exposure through an infrared filter. This is based upon the disclosures of 
A. Murray of the Eastman staff that most pigments used in paints are trans- 
parent to or highly reflective of the infrared rays. Other discussions of the 
subject are contained in articles by F. J. Tritton {Phot. /., Vol. 78 (1938), 
p. 732); H. H. Lerner and W. Perelstrus (Phot. Tech., Vol. 1 (1939), November, 
p. 25, December, p. 35); and J. S. Friedman (Am. Phot., Vol. 30 (1936), 
pp. 692, 771, 837; Vol. 31 (1937), pp. 129, 584; Vol. 33 (1939), PP- 372, 928), 
from which this historical abstract has been compiled. 

The necessity for masking arises from the poor spectral characteristics of 
the colors used in the synthesis processes. The red filter separation is a record 
of the red densities as they are present in the original. The positive made 
from this must deposit red densities at every point of the reproduction, to 
the extent determined by the relative transmissions at each point of the red- 
filter negative. This is the sole function of the red filter or cyan positive. 
But the cyan colors available to the color technicians are extremely poor in 
their green and blue transmissions. Instead of transmitting 100 per cent 
of the blue and green rays, the average transmissions will be in the neighbor- 
hood of 45 per cent. Therefore the cyan image will deposit not only red densi- 
ties, but also blue and green densities. In the same manner, the magenta 



MASKING 275 

colors suffer as far as the blue transmissions are concerned. It is only the 
yellow pigments which have a sufficiently high transmission of the desired 
primaries to be practically perfect. The average magenta will transmit up- 
wards of 85 per cent of the red rays that are incident upon it, but only 45 
per cent of the blue. Let us take these figures for our general average. 

Consider a point on the original that contains a color that reflects 75 per 
cent of the total blue rays which fall upon it, 30 per cent of the green, and 
only 3 per cent of the red. We can convert these transmissions into den- 
sities by use of the formula Density = 2 — log (% Transmission). The blue 
therefore can be assigned a density value of 0.13, the green a value of 0.52, 
and the red a value of 1.52. If the processing of the negatives and the posi- 
tives has been done accurately, the magenta positive which deposits the 
green densities will have a value of 0.52, the cyan a value of 1.52, and the 
yellow a value of 0.13, after they have been converted into color. But now 
the effects of poor transmissions of the colors to light of their own ranges 
become evident. Suppose that the cyan dye transmits but 45 per cent of the 
green and blue. When this dye is present in sufficient quantity to give a 
density equal to 2.00, there will be a considerable amount of blue and green 
light absorbed, with consequent deposition of considerable quantities of blue 
and green densities. Since at maximum concentration but 45 per cent of the 
green and the blue is transmitted, it becomes a simple matter to calculate 
the amount of undesirable density that is deposited. To obtain an accurate 
determination, it is only necessary to measure the densities of a heavy cyan 
image through all three filters. Let us suppose that when the value through 
a red .filter is 1.55, the value through the green filter is 0.31, and the value 
through the blue filter also 0.31. 

Theoretically the values should have been 1.55 through the red, and zero 
through the other two filters. Therefore at each point of the cyan image, 
there have been deposited green and blue densities equal to 20 per cent of the 
red. These deposits bear no relationship whatsoever to the requirements of 
the picture, because the true green densities will be put down by the magenta 
image. Since the transmissions of the blue and green are 45 per cent of 
theory, and a transmission of 45 per cent is equivalent to a density of 0.35, 
this will be the maximum amount of undesired densities that the cyan dye 
could deposit. Applying these results to the point in question, we see that 
the cyan dye will deposit these densities 



Red 


1.52 


Green 


0.30 (= 20% of 1.52) 


Blue 


0.30 



The magenta dye, transmitting practically 100 per cent of the red rays, will 
deposit no red densities. But it transmits only 45 per cent of the blue rays, 
hence it will deposit a considerable quantity of blue densities. Since the 



276 HISTORY OF COLOR PHOTOGRAPHY 

inefficiency of the magenta to the blue is the same as the inefficiency of the 
cyan to this primary, we can assume that here also 20 per cent will be de- 
posited. Hence the magenta dye will give these densities: 



Red 


0.00 


Green 


0.52 


Blue 


0.10 (= 20% of 0.52) 



The yellow image, being practically perfect, will deposit merely blue densi- 
ties, and these to a value of 0.13. The final color will have the composition, 
in terms of densities: 

Red 1.52 + 0.00 + 0.00 = 1.52 

Green 0.30 + 0.52 + 0.00 = 0.82 

Blue 0.30 + 0.10 + 0.13 = 0.53 

These densities will correspond to the following transmissions: 



Red 


3% 


Green 


iS% 


Blue 


30% 



The color is still a blue, but now it has a very pronounced bluish-green tint, 
and is extremely dull. The original color was an extremely bright blue, with 
but a slight greenish tint. The shade, therefore, has been made considerably 
greener, and the saturation reduced to a very low value. 

This degradation of color can be corrected if some means can be found to 
reduce the deposition of blue densities by the yellow printer to an amount 
equal to 20 per cent of the red densities at each point, and an added amount 
equal to 20 per cent of the green densities at each point. Apparently 0. Pfen- 
ninger and E. C. Townsend (see above) must be given the credit for first 
disclosing how this can be done. 

Let us suppose that our negatives have been developed to a gamma of one. 
The cyan printer is made from the red-filter negative, and it deposits the red 
densities. For all practical purposes, this printer is the only one which de- 
posits red densities, since the magenta and the yellow inks or dyes transmit 
the- red with great efficiency. Now if another positive is made of this printer, 
but developed to a contrast equal to 20 per cent of the contrast of the first, 
each point on this positive will have a density which is 20 per cent of the 
density of the corresponding point in the true cyan printer. These are the 
densities that must be subtracted from the yellow printer. The subtraction 
is carried out by registering the weak positive with the yellow printer or blue 
filter negative. Center our attention again at the point where the red density 
is 1.52 and the blue density is 0.13. The corresponding point in the blue filter 
negative has a very heavy silver deposit, since but a very thin deposit is re- 
quired in the positive. Let us suppose that the value of this deposit is 1.75 
in the blue-filter negative. The true cyan positive has a value of 1.52 at the 



MASKING 277 

same point, hence the mask has a value of 20 per cent of 1.52 or 0.30. The 
combined mask and blue-filter negative has a value of 1.75 + 0.30 or 2.05 
at this point, and when the yellow printer is exposed through such a masked 
negative, and developed to a gamma of one, there will be a corresponding re- 
duction in the image at this point. Therefore we have succeeded in sub- 
tracting the desired densities from every point of the yellow printer. 

However, there is one point to be brought out. The original color calls for 
a deposition of a blue density equal to 0.13. We have added a density equal 
to 0.30 to this point on the blue-filter negative. Our exposure is such that 
without the inclusion of the mask there would be formed a density of 0.13, 
so that with the mask added, the exposure would be insufficient to print through 
the added density. If more exposure is given there will be distortion all along 
the line. Therefore this is no solution. We conclude, therefore, that masking 
can correct only when the color densities are sufficiently high, so that they are 
greater than the densities deposited by the inefficient transmissions of the 
dyes themselves. In our present example, the cyan, which is the principal 
color at this point, will deposit a blue density of 0.30. The magenta, which 
should shade the cyan so that it becomes somewhat less green, also gives a 
deposit of blue to the value of 0.10. It is therefore impossible for us to ob- 
tain a value for the blue density which is less than 0.40. This much falsifica- 
tion and degradation of colors must be accepted. 

When a negative having a contrast of 1.00 is registered with a positive whose 
contrast is 0.20, the result will be a negative whose contrast is 0.80. This is 
evident from the following example. Included in the picture is a ten-step 
gray scale whose steps have a constant differential of 0.15. If the lightest 
step has a value of 0.30, the second will have a value of 0.45, etc. The gray 
scale will be reproduced in each of the three separations. The blue-filter 
negative, if developed to a gamma of 1.00, will have values for the different 
densities as follows: 



Step 


Density 


Step 


Densit 


1 


0.30 


6 


1.07 


2 


0.4s 


7 


1.20 


3 


0.61 


8 


I-3S 


4 


0-7S 


9 


1.49 


5 


0.90 


10 


1.60 



The other negatives, also developed to the same gamma, will have values 
very close to this. The original scale had the following values: 



Step 


Density 


Step 


Density 


1 


1.65 


6 


0.90 


2 


1.50 


7 


o.75 


3 


i.35 


8 


0.60 


4 


1.20 


9 


o.45 


5 


1.05 


10 


0.30 



278 HISTORY OF COLOR PHOTOGRAPHY 

The positive made from the red-filter negative has densities: 



Step 


Density 


Step 


Density 


i 


i.6o 


6 


0.89 


2 


i.5o 


7 


0.76 


3 


1-35 


8 


0.61 


4 


1.20 


9 


0.47 


5 


1. 05 


10 


o.35 



The weak positive made from the red-filter negative has densities that are 
only 20 per cent of these, thus: 

Step Density Step Density 





1 0.32 


6 


0.18 




2 0.30 

3 0-27 

4 0.24 

5 0.21 


7 
8 

9 
10 


0.15 
0.12 
0.09 
0.07 


\ combined mask and yellow negative, will have values: 


Step No. 
1 
2 
3 
4 
5 


0.30 + 0.32 = 0.62 
0.45 + 0.30 = 0.75 
0.61 + 0.27 = 0.88 
0.75 + 0.24 = 0.99 
0.90 + 0.21 = 1. 11 


Step No. 
6 

7 
8 

9 
10 


1.07 + 0.18 = 1.25 
1.20 + 0.15 = 1.35 
1.35 + 0.12 = 1.47 
1.49 + 0.09 = 1.58 
1.60 + 0.07 = 1.67 



This, when plotted against the original (Fig. 106), is seen to have a gamma 
of 0.80. 

This can be made into a general rule. The combination of a negative with 
gamma a, with a positive whose gamma is b> will be a negative with a gamma 
a — b, if b is less than a. If it is greater, then the result will be a positive 
with a gamma equal to b — a. Since it is essential to preserve equality of 
gammas in the three printers, the use of a mask on one negative will necessitate 
the extra development of the positive through that masked negative, in order 
to maintain balance. Therefore, if a 20 per cent mask is in order, there will 

be required a development to a gamma -^— > or 1.25. 

There is another method for maintaining this balance. The operator knows 
beforehand that he is to deliver a set of masked negatives, and that the ma- 
genta printer is to receive a single mask from the cyan negative equal to 20 
per cent of the cyan printer, and the yellow negative is to receive a mask 
from both the cyan and the magenta printers equal to 20 per cent of the 
densities deposited by those printers. He develops the red-filter negative 
to a contrast of 1.00. The green-filter negative is to be developed to that 
contrast which, when combined with the contrast of the cyan mask, will 
yield an overall contrast of 1.00, at the same time preserving the 1:5 ratio 



MASKING 279 

between the contrasts of the green-filter negative and the cyan mask. If % 
is the value of the contrast of the mask, then 5# is the contrast of the green- 
filter negative. We can set up the equation $x — x = 1.00 which stipulates 
that the differences in the contrasts of the green-filter negative and cyan mask 
must be 1.00. Solving, we get x = 0.25, so that the mask must be developed 
to a contrast of 0.25, while the green-filter negative must be developed to a 
contrast five times that, or 1.25. When the two are combined we will have a 



1.70 


:. 
























.* 






















^f* 










































*» 




J/^ 








* uo 


























£ 1 °,80 












s 






























































5 § oso 






























■ 








































































S 4. 




























14 


50 


1. 


so 


1. 


20 





.90 


0- 


60 


0. 


30 



FIG. 106 



masked negative whose overall contrast is 1.00, and where the ratio between 
the contrast of the mask and negative to be masked is retained. 

The other negative (blue-filter) is to be corrected by registering two masks 
each 20 per cent of the respective printers. The contrast of the two masks 
in combination will be 040. There is now a ratio of 0.40 to 1.00 which must 
be preserved. Since the two masks are equal in contrast they may be con- 
sidered as one. If x is the contrast of the combined masks, then 2.5a; is the 
contrast of the negative. Therefore the equation takes the form: 

2.50a; — # = 1.00 

1.5a; = 1. 00 

x = 0.67 

Under these conditions, the contrast of the masks must be each 0.33 and 
the contrast of the blue-filter negative 1.67. 



2 8o HISTORY OF COLOR PHOTOGRAPHY 

This result can also be obtained in another manner which does not call for 
such drastic measures. Since masking is confined today practically com- 
pletely to still life or to separations made from copy, speed is of no value. 
The yellow printer is to be masked by equal red- and green-filter masks. There- 
fore, if a special negative is made, either by giving a plate exposures through 
both the red and green filters, or by exposing through a minus blue (yellow), 
there will be obtained a negative combining the red and green densities. 
This can also serve as the black-printer negative. The mask for the yellow 
printer is made from this negative and it is developed to an overall contrast 
that is equal to two-fifths the contrast of the blue-filter negative. The printer 
made from this is developed to an overall contrast of 1.67. Thus if the original 
blue-filter negative has a contrast of 1.50, not an impossible value except with 
the fast negative materials, the mask will be developed to a contrast of 0.60, 
and the yellow printer to a contrast of 1.10, so that the overall contrast is 
(1,50 — 0.60) X 1. 10, or 1.00. 

The necessity for masking in the photomechanical trade arose from the 
need for eliminating the very expensive and time-consuming operations of re- 
etching and hand retouching. Not only are these operations costly, but they 
give a mechanical appearance to the picture, removing all such aspects as are 
now grouped together under the term photographic quality. When a study 
was made of the properties of the printing inks and the deficiencies of the 
spectral characteristics noted, the idea of photographic re-etching or sub- 
traction of densities by use of masks became immediately apparent. For a 
short space of time the photoengraving industry, never noted for scientific 
approach to its problems, went overboard for this technique without inquiring 
too closely as to possible discrepancies between working conditions and condi- 
tions under which masking is possible. The net result was that in a very 
short time it became evident that although masking helped, it did not by 
any means eliminate the re-etching. This caused a reaction to set in, and far 
too many of the shops gave up the practice and returned to the use of their 
good right eyes. But the brief introduction of the technique had one good 
effect. It made the shop technicians realize the value of densitometric control. 

From the surface, it would appear that masking should be an excellent 
corrective technique to use under any conditions. But the failure obtained 
in the industry cannot be lightly dismissed. Let us examine what conditions 
must be preserved in order to be able to apply a masking technique intelli- 
gently. First of all, the masks and the negatives to which they are applied, 
are both continuous in tone. When a correction is to be applied at a given 
point, it is because at the same point there has been or will be deposited some 
unwanted densities. This is of course axiomatic in sub tractive processes — 
and masking can be applied only to subtractive processes. 

Now consider the photomechanical product. When the halftone negatives 
are made, the different separations are angled differently with respect to the 
screen so that the imposition of the three screen patterns, one on top of 



MASKING 281 

the other, will not cause an annoying moire pattern to form. This solved the 
moire problem, but it created another which is seldom mentioned but which 
probably explains the existence of the most of the headaches in that industry. 
This problem is that it separates the undivided color dots so that in practically 
all areas with the exception of the blacks and the very dark colors, the color 
dots fall side by side, rather than one on top of the other. An examination 
of the various tones in a magazine illustration under a moderately powered 
glass, will quickly show what is meant. In the highlights and the lower middle 
tones, we do not operate under a subtractive system, but under an additive 
one. This, more than any other reason, explains why the cyan and the magenta 
approach more nearly the green and red of the two-color processes than the 
cyan and magenta of a three-color, process. 

In the middle tones, the sizes of the dots become sufficiently large to oc- 
casion some overlapping. In this region, which extends to the region of the 
deep grays and the dark colors, the system is operating partly additive and 
partly subtractive. It is only in the deepest shadows that it acts subtractively, 
and even here due to the common practice of maintaining a dot structure, 
there will be a slight region where the additive principle is working. The 
finer the screen, the greater the mixture between the two, for then the additive 
portions will blend more smoothly. It must be recalled that in tone values 
the halftone system is strictly additive in operation. Grays are obtained by 
mixing white with black (regions of pure white are mixed additively with 
regions of pure black). Under these conditions it is no small wonder that 
masking fails to eliminate all re-etching and, depending upon the subject 
matter, may not even eliminate a large proportion of it. This explains the 
claims of some technicians that masking eliminates from fifty to seventy-five 
per cent of the hand work, and of others, who state that it accomplishes very 
little. 

Although of dubious value in the photomechanical industry, masking properly 
applied can play a great role in the field of photographic reproduction. Here 
there are no complicating problems of tone separation. The different color 
densities of corresponding points fall directly one on top of the other. The 
application to photography can be traced back at least to 191 5, when two 
patents were issued to W. F. Fox (U.S.P. 11 661 21 and 11 66 12 2). Here was 
disclosed a method for increasing the color contrasts in a two-color process. 
The negative of one color separation is united with a positive of another, to 
make a masked printing negative. In his all-inclusive patent (Re-issue 18680) 
Dr. Troland also utilized masking, but for an entirely new and novel purpose. 
It is to be recalled (Chapter on Monopacks) that the front element of a Tro- 
land or Friedman bipack, was a stratified emulsion. The lower half contained 
the blue densities, and the upper half, the red. Between the two, and acting 
as an opaque barrier between them, were the unfixed silver halide salts. 

A copy of the red densities is made by reflection printing of the upper image. 
The copy is made so that the size is identical with that of the original, and the 



282 HISTORY OF COLOR PHOTOGRAPHY 

contrast is made equal to that of the red-density image. When these are 
registered with each other, the positive reflection copy just neutralizes the 
negative of the red densities. If this combination is printed, only the blue 
densities will modulate the exposing light. In this way the blue densities are 
obtained. This technique would yield separations of a quality much superior 
to those obtained by copying. 

This type of masking technique has been suggested by J. S. Friedman, 
(p. 274) as a means of greatly increasing the lens speed of a one-shot camera. 
In order to achieve this, the niters in a one-shot camera are replaced by an A, 
a Wratten No. 12, and a Wratten No. 32 respectively. Behind each of these 
filters is placed a panchromatic emulsion. The image behind the A filter will 
represent the red densities. The image behind the yellow filter will represent 
the green and the red densities, and to an equal extent. If the contrast of the 
No. 12 negative be 1.50, then the mask made from the A filter should have 
an overall contrast of 0.75. The combination of the two will be a negative 
having a contrast of 0.75. The mask will subtract red densities from the 
No. 12 negative to a contrast of 0.75, but the red densities present in this 
negative have a contrast that is equal to 0.75. Hence the masking positive 
will just neutralize the red densities in the No. 12 negative, leaving only the 
green densities to come through. In a similar way the combination of a mask 
from the A filter negative, developed to an overall gamma equal to one- 
half that of the contrast of the No. 32 filter negative, with that negative will 
neutralize the red densities here, and allow the blue to come through. By this 
technique there is obtained, after masking, the three color separations. 

The filter factors of the Nos. 12 and 32 filters are 2. The filter factor for 
the A filter is also 2 (for light of Mazda quality, 3200 K). We have, therefore, 
to divide the light intensity entering the camera lens equally into three parts. 
If this is done, and if emulsions with a Weston rating of 50 are used, there will 
be obtained speeds of approximately Weston 9. This measures quite favorably 
with the speeds of 1 to 2 that would be obtained under similar conditions, 
using the more normal technique. 

The Eastman Kodak Company, in their booklet describing the Wash-Off 
Relief Process, describe a method for masking that can be generalized to all 
other types of processes. This scheme is also discussed in United States patent 
2169009 issued to M. W. Seymour of the Eastman Kodak Co., and in English 
patent 517020, issued to R. F. W. Selman, of Kodak, Ltd., England. We 
will first discuss the special scheme outlined in the booklet, then the generaliza- 
tion. The Wash-OfT Relief dyes are very efficient in their red transmissions, 
hence the red-filter negative need not be masked. This negative is printed 
by contact upon Wash-Off Relief material, and processed to form the cyan 
matrix which is dyed up to the extent required for the final cyan image. The 
magenta matrix must be corrected for the green densities deposited by the 
cyan image. 

The cyan-dyed matrix is placed in .registry with, the green-filter negative, 



MASKING 283 

and the combination is printed by contact upon Wash-Off Relief material* 
The automatic character of the masking now becomes evident. The masked 
negative should be printed with green light. Therefore the intensity of the 
light at any one point is first reduced to the extent to which the cyan image 
deposits green densities at that point, and then is modulated by the green- 
filter negative lying below the mask. This would require that the Wash-Off 
Relief emulsion be green-sensitive, which it is not. But the amount of green 
masking is approximately the same as the blue, so that but very little dif- 
ference would be obtained if the printing of the magenta matrix were done 
with blue light instead of green. This is taken care of by the emulsion itself, 
since it is truly blind. The magenta matrix is processed to completion, and 
is dyed to the exact extent that would be used in the final image. The two 
matrices, dyed cyan and magenta, are combined and the combination placed 
in registry with the blue-filter negative. Now a contact print made from this 
combination will be correctly masked for the poor blue transmissions of both 
the cyan and magenta dyes actually used, because the exposing light must 
first pass through the cyan and magenta positive images before passing 
through the blue-filter negative. Therefore the light is modulated, first of all 
to the extent of the blue densities deposited by the dyes, then by the blue- 
density negative. 

As disclosed by the Eastman technicians, this semi-automatic scheme is 
useful only in the case where same-size images are made. The scheme is 
termed semi-automatic because the green densities are corrected by the blue 
absorptions of the cyan dye, rather than by the green absorptions. But as 
noted above, the two approximate each other very closely. It is possible to 
generalize the technique to make it useful for enlargement. The masks, in 
this case, are made by contact and registered with the negatives in the manner 
disclosed above. But these are not used for the final print. New matrices 
are made by projection through the masked negative, so that enlargements 
to any desired degree can be obtained. 

Should one use other methods of making the positive prints, a similar type 
of technique could be evolved. Let us suppose that the print is to be made by 
color development. The red-filter negative is first printed by contact upon 
a film, and this is converted into a cyan image by coupler development (cf. 
chapters on Color Development for details). The colored image is registered 
with the green-filter negative, and the magenta image printed by contact 
from this combination, etc. With different Systems, it may be advantageous 
to mask the red-filter negative, as well as the others. This is a rather obvious 
extension of the present technique. 

The current popularity of the monopack films typified by Kodachrome, 
bids well to extend the idea of masking to still another field, the making of 
separations from monopack transparencies. Here more than anywhere else, 
a masking technique is desirable. This matter is discussed rather critically 
by J. S. Friedman in several of his columns in American Photography (cf. I.e., 



284 HISTORY OF COLOR PHOTOGRAPHY 

and Vol. 34 (1940) p. 128). It also forms the subject matter of a patent by 
Dr. Bela Gaspar (Eng. P. 494341). A colored image in a transparency consists 
of three dye part-images lying in three superimposed emulsion layers. One 
image is magenta in color, hence gives the green densities. Another image 
is cyan, while the third image is yellow. These give the red and blue densities 
respectively. 

When the composite is illuminated by means of red light, the rays will pass 
completely undisturbed through the magenta and yellow layers, provided 
these are theoretically correct. Upon passing through the cyan, absorption 
will take place to correspond to the image density. Thus the red light becomes 
modulated in accordance with the amount of red densities present in the trans- 
parency. In a similar manner it is possible to obtain the green and the blue 
densities, by exposing with green and blue light respectively. Provided the 
dyes in the transparency are accurately chosen, it becomes possible to obtain 
excellent color separations from this type of color transparency. But no 
dyes are known that are perfectly correct in their absorptions and transmis- 
sions. Every magenta dye absorbs some red and blue. Hence when the 
red light passes through the magenta image, some of the red rays become 
absorbed. The light becomes modulated to correspond to the green densities 
to the extent to which absorption takes place. Therefore there is imposed 
upon the record of the red densities a weak record of the green densities and 
the quality of the separation becomes materially reduced. A similar story 
is true with respect to the green and the blue densities, and unfortunately, to 
a considerably greater extent. Masking, therefore, is in order. 

If the amount of correction is known, it will be possible to subtract this 
amount from the negatives, by adding positives from the other negatives to 
each separation. This would be the normal manner of procedure, correspond- 
ing to the technique disclosed above. Dr. Gaspar, however, adopted a more 
suitable method. In general, the red-filter separation need not be corrected. 
This is made by contact printing with red light, and the negative is processed 
to yield a cyan dye image whose gamma, measured through a red filter, is 1.00. 
After drying, the colored negative image is registered with the original trans- 
parency. Where this is done, the cyan positive image in the transparency 
becomes completely neutralized by the cyan negative mask and, in so far as 
the exposing green light is concerned, the transparency has overlayed upon 
it a uniform cyan fog. Just as much green light will be restrained in the 
shadows as in the highlights. Now if an exposure be made with green light, 
the light will be modulated correctly by the magenta image, and uniformly 
reduced in intensity by the cyan image plus mask. The efficiency of the 
yellow to green light is quite high, so but secondary or tertiary effects will be 
noted here. 

To correct for the blue it is necessary to use two masks, one a cyan made as 
above, and the other made from the cyan-masked transparency by means 
of green light. This last yields the magenta mask, hence it must be colored 



MASKING 285 

a magenta. The combined cyan and magenta masks, registered with the 
original transparency, are now used to make the blue-filter separation. 

This method when applied to Kodachrome transparencies, gives automatic 
correction provided the colors used to dye the negative masks are the same 
as those used in the original Kodachrome. Dr. Gaspar proposed his scheme 
for use with Gasparcolor, and the materials upon which the masks were made 
could be identical with those used in Gasparcolor. But the Kodachrome dyes 
are not known. The patents list a number of different magentas, cyans, and 
yellows. However, no tremendous differences lie between the various colors. 
A good set to use would be para-nitro-phenyl-aceto-nitrile for the magenta, 
di-chlor-alpha naphthol for the cyan, and aceto-acet-anilide for the yellow. 
These are the coupling agents that are added to a developer containing 2-amino- 
5-di-ethyl-amino-toluene-mono-hydrochloride. A complete discussion of color 
development is contained in Chapter 23, but since this procedure is of in- 
creasing importance, the modus operandi will be discussed now. 

Coupler developers yield dye images together with the silver, so that after 
the removal of the silver, a pure dye is left behind. The developer is con- 
cocted by mixing three stock solutions. 



Stock Solution A . 




Water 


1000 parts 


Potassium metabisulphite 


10 parts 


2-amino-5-di-ethyl-amino-toluene-mono-hydrochloride 5 parts 


Potassium bromide 


10 parts 


Potassium thiocyanate 


5 parts 


Stock Solution B. 




Water 


1000 parts 


Sodium carbonate 


100 parts 


Stock Solution C (for magenta). 




Para-nitro-phenyl-aceto-nitrile 


1 part 


Acetone 


100 parts 


Stock Solution C (for cyan). 




Di-chlor-alpha-naphthol 


2 parts 


Acetone 


100 parts 


Stock Solution C (for yellow). 




Aceto-acet-di-chlor-anilide 


10 parts 


Acetone 


100 parts 



For use, take 100 parts of solution A, 100 parts of solution B, 5 parts of 
solution C, then' make up with water to 1000 parts. Develop for three minutes 
at 70 F. Each exposure must be developed in a fresh solution, which should 
be immediately discarded after use, as the solutions do not keep except in 
the stock form described above. After development the images are fixed in 
hypo to which a little metabisulphite has been added. Care must be taken 



286 HISTORY OF COLOR PHOTOGRAPHY 

that the acidity is maintained at a minimum, as acids destroy the dye image. 
A thorough wash removes the hypo. Treatment with one per cent ferricyanide 
plus one per cent bromide converts the silver to silver bromide, which can then 
be removed with hypo. After a wash, the images can be dried and registered 
with the original. It should be noted that this type of masking is analogous 
to the masking described by the Eastman Kodak Company, and described 
above. But there the masking is accomplished by registering the dyed matrix 
with the separation negatives, while here it is achieved by registering a colored 
separation with the original copy. This type of masking allows separations to 
be color-corrected for the deficiencies of the dyes in the original while making 
the separations, after which the ordinary type of masking can be applied to 
correct for the deficiencies of the synthesis colors. 

It is not essential, of course, to use color masks if the amount of correction 
to be applied is known. In that case, it is sufficient to prepare black-and- 
white masks of the proper intensity, register these with the original, and make 
the separations through the combined black-and-white masks and colored 
transparency. 

A scheme along these lines is outlined in the Gevaert " Graphic Art Hand- 
book," where it is proposed as a dodge to be used to reduce the excessive 
contrast in Kodachromes. Instead of making three separations, three masks 
are first made through filters that are complementary to the separation filters. 
Thus the mask for the blue-filter separation, is made through a yellow (Wratten 
Nos. 9, 12, 15, or 16). The mask for the red-filter separation is made through 
a cyan (Wratten Nos. 43, 44, 44A, 45, 64, 65, or 65 A).' .The mask for the green- 
filter separation is made through a magenta filter (Wratten Nos. 31, 32, 33, 
34, or 35). The contrast to which the masks are developed, is determined by 
trial and error. Upon completion, the mask made through the cyan filter is 
registered with the original, and the combination exposed through the red 
filter to yield a cyan printer negative. The other two separations are made 
in an analogous manner, the masks being changed before each exposure. 

In his discussion of this procedure, J. S. Friedman (I.e.) suggested that it 
-is possible to combine the Gaspar, with the Gevaert schemes. Instead of using 
a yellow filter, the mask can be exposed partially through a red filter, and 
partially through a green, the proportion of the two exposures being in the 
inverse ratio to the sensitivity of the emulsion to these colors. In that way 
it. is possible to obtain equivalent red and green contributions to the mask. 
By this means, there have been copied upon the same mask the images in the 
cyan and magenta layers of the original. The mask is color-developed a 
magenta as disclosed above. It is fixed, washed, and treated with ferricyanide- 
bromide, but instead of being fixed, it is thoroughly washed to remove the 
last traces of bleach solution, then color-developed a cyan. By this means 
there will be prepared a blue-colored mask, which, after the silver has been 
removed, is registered with the original Kodachrome. When light is trans- 
mitted through this combination, the yellow image will modulate the blue 



MASKING 287 

rays and allow the other two colors to pass through unhindered. The cyan 
positive in the Kodachrome will be completely neutralized by the cyan nega- 
tive image in the mask, so no modulation of these rays will take place. Similarly 
the magenta image in the Kodachrome will be neutralized by the magenta 
positive image in the mask. Therefore, the light, after passing through the 
Kodachrome will represent the image present only in the yellow layer of 
the Kodachrome. Since the red and green rays pass unmodulated, though 
reduced in intensity, it is necessary to make the exposure through a blue' 
filter, to avoid loss of brilliance and formation of excessive over-all ground 
density. 

This appears to be a rather long-drawn-out procedure to accomplish a 
desired result. It is possible, of course, to mix the cyan and magenta couplers 
in the proper ratio, so that equivalent amounts of cyan and magenta dyes are 
formed in the first development. Or, if the degree of absorption of blue by 
the cyan and magenta be known, the mask could be developed to that extent, 
as a black and white. A still simpler, though slightly less thorough way, 
would be to process the plate a magenta, or a cyan, and let it go at that. 
The green-filter separation is to be masked by means of a magenta filter or 
by partial exposures through the blue and red filters. Theoretically, this 
should be converted into a green by color development, but a more practical 
practice would be to leave it cyan or yellow in color. Similarly, the mask 
for the red filter separation is made through a cyan filter or by partial ex- 
posures through the blue and the green filters. This mask should be processed 
to yield a red. The simplified technique would be to leave it magenta or 
yellow in color. This type of masking is termed secondary, since each mask 
is made by exposure to secondary colors rather than to primary. 

Since the colors used in the original Kodachrome are quite deficient, much 
better results would be obtained were it possible to mask these colors during 
the preparation of the original. A scheme to accomplish this is disclosed by 
N. E. Brookes (Eng. P. 496997). It is to be recalled that the Kodachrome 
transparency contains three layers, the bottom of which contains the cyan 
image, the central layer the magenta image, and the top contains the yellow. 
After the original exposure, the film is developed, the silver image is removed, 
and the remaining silver halides developed a cyan color. A bleach destroys 
the dye in the two upper layers, and converts the silver to silver halide, which 
is then developed to a magenta. This process is repeated, this time the 
bleach being confined to the top layer, which is developed a yellow. 

This is the normal procedure. The Brookes modification is to expose the 
reformed silver halide salts in the second and third layers, by means of green 
light, through the already formed cyan image. Therefore the intensity of 
the latent image produced in the magenta- or green-density layer, will be 
reduced to the extent that green is absorbed by the bottom image. This is 
developed a magenta, and the procedure is repeated, this time using blue 
light and exposing through the images already formed. Thus the intensity 



288 HISTORY OF COLOR PHOTOGRAPHY 

of the latent image produced in the top layer is modified to the extent that 
blue is absorbed by the magenta and cyan images in the other layers. 

In general, the practice of masking involved the use of several plates or 
films, registered with respect to each other. Special cameras were suggested, 
as pointed out above, which enabled one to register the mask with an unex- 
posed negative plate, so that the final result would be a single emulsion layer. 
There have been other suggestions that accomplished the same result. One 
of them has been discussed in detail, and involved the use of negative masks 
registered with the original, if that is a transparency. The I.G. (Ger. P. 
593666) suggested the following scheme: The separations were processed to 
retain the gelatin in a soft condition. The final step consisted in bathing them 
with a light-sensitive dichromate solution. Since the red-filter separation re- 
quires no correction, this is not so treated. The green-filter separation, after 
sensitization with dichromate, is registered with a red-filter positive, and 
exposed to arc light. Where the light is incident upon the green-filter negative, 
tanning of the gelatin takes place. After exposure the excess dichromate is 
washed out, and the plate dyed with a black pinatype dye. Since this is not 
absorbed by tanned gelatin, a positive is formed, thus creating the mask in 
the same emulsion layer as the negative. The contrast of the positive is 
governed by the acidity of the dye bath, giving the operation a wide latitude. 

The Interchemical Corporation (Eng. P. 501 661) would use a double-coated 
emulsion upon which to make the exposures. One of the emulsions is a nega- 
tive material, the other a positive. The plates are not fixed out, with the 
exception of the red-filter separations, but are registered with the red separa- 
tion, and a positive printed upon the second emulsion layer. A rather novel 
scheme, is the one suggested by L. T. Troland of the Technicolor Motion 
Picture Corporation (U.S. P. 2098441 and 2098442). Use is made of the 
Herschel effect, where the intensity of a latent image becomes reduced by 
exposure to red or infrared light. The procedure is applied to the printing 
of the matrix. This is first normally exposed to the negative under considera- 
tion. It is then exposed through the proper mask by means of red light. The 
loss of latent image will correspond to the mask. Other techniques of masking 
are disclosed by E. A. Weaver (U.S.P. 2183598 and 2186053) and by J. A. C. 
Yule (U.S.P. 2176518). A tremendous number of patents in this field are 
applicable only to the photomechanical trades. Of these, only a very few 
received consideration in this chapter. 

It was stated above that masking has its real appeal in the making of sepa- 
rations from color transparencies such as Ansco Color or Kodachrome. A 
rather naive treatment of the subject was outlined. At this point we will re- 
consider the problem from a more sophisticated point of view. 

The usual procedure used in making color reproductions, be they purely 
photographic or photomechanical in nature, is to copy the original through 
the three separation filters. These, as was pointed out in Chapter 3, have 
mutually-exclusive transmissions, so that they have no interdependence what- 



MASKING 289 

soever. When a red-filter copy is made, the image produced can be represented 
mathematically by the form 

a\r + o.oog + 0.006. 

This means that the photograph is a graphical representation of the red 
densities to the value of a±. If the photograph is a negative, then the constant 
a\ is written with a negative sign in front, so that in the case above* since the 
red separation is a negative of the original, we should write 

— a,\r + o.oog + 0,006. 

Let us make another simplification. We will focus our attention upon one 
point in the original, a point that has a density of 1.00 when measured through 
each of the three filters. Then the value a\ or — a\ will be the contrast to which 
the negative was developed. In a similar manner we can write for the green 
and blue-filter separations, the forms 

o.oor — 02g + 0.006, 
and 

o.oor + o.oog — a 3 6. 

Taken as a unit, they give rise to the mathematical " color matrix" whose 
elements are the coefficients in the three equations, thus 



- a x 











-a* 





. 





— a* t 



For a truly balanced set of separations ai, 02 and a 3 must have the same value. 
Thus the ideal balanced set of separations can be represented by a color matrix 
such as 



(—a o o\ 

o — a o J 

o o — a) 



Now suppose we photograph the same object by means of a film like Ansco 
Color or Kodachrome. We obtain a color image that is a reproduction of the 
original, which is composed of three monochrome images superimposed upon 
each other. Each monochrome is supposedly the image of a single primary. 
Thus the cyan image gives the pattern of the red densities, the magenta gives 
the pattern of the green densities, and the yellow the pattern of the blue 
densities as they have been registered on the transparency. Such a transparency 
can serve as a new original for the purposes of reproduction, provided it is 
possible to isolate each of the individual part images. The beauty of such a 
procedure would be that it would eliminate the expensive, intricate, and ex- 
ceedingly delicate one-shot camera, and project the problem of separations 
into the precision laboratory where it belongs. 

In order to make separations, we photograph the transparency through the 
three separation filters. If the absorptions of the cyan were ideal, i.e. if the 



2 9 o HISTORY OF COLOR PHOTOGRAPHY 

cyan absorbed no blue or green light, and if the other two colors absorbed only 
the light they were meant to absorb, this scheme would work perfectly, and 
one-shot cameras would immediately be relegated to the scrap heap. But the 
colors are all impure, and in general absorb not only the primary they were 
intended to absorb, but also the others. A typical set of secondaries or sub- 
tractive primaries is the following 

cyan = i.oor + 0.50^ + 0.306 

magenta = 0.15/* + i.oog + 0.456 

yellow = o.o5r + o.iog + 1.006 

The cyan dye, when it is present in sufficient quantity to yield a density of 
1. 00 through the red filter, yields a density of 0.50 through the green, and a 
density of 0.30 through the blue filter. Similarly, when the magenta is present 
in a quantity sufficient to give unit density through the green, it also gives a 
density of 0.15 through the red and 0.45 through the blue. When the yellow is 
sufficiently strong to give a density of 1.00 through the blue, it also gives a 
density of 0.05 through the red, and 0.10 through the green filter. 

A color transparency is processed in such a manner that a gray scale is 
reproduced as a gray, and presumedly a density of 1.00 in the original is re- 
produced as a density of 1.00. This is the goal, and the departure from it is 
slight, so it will represent second-order corrections. We will therefore ignore 
them in this analysis. Given a reproduction process ' using the secondaries 
noted above, there is one, and only one, combination of the three which will 
yield a density of 1.00 through the three filters. Let us suppose this is the 
combination which utilizes a concentration x of the cyan, y of magenta, and 
2 of yellow. This is exposed and processed to yield a density of 1.00 through 
each filter. 

Consider the red filter. The cyan will contribute an amount — i.oo#, the 
magenta will contribute an amount — 0.15^, and the yellow — 0.052. To- 
gether the total density will be — 1.00, the negative signs indicating that the 
image obtained will be a negative. A similar result will be obtained from the 
green and blue-filter separations, so we have the three equations 

— i.oo# — o.i5;y — 0.052 = — 1.00 

— 0.502 — i.ooy — 0.102 = — 1. 00 

— 0.302 — 0.45^ — 1.002 = — 1. 00 

Solving for x, y, and 2, we obtain x = .90, y = .50 and 2 = .50. Substituting 
this in our equations, we get for the dyes as they are actually present in the 
film for a density of 1.00 

cyan = o.9oor + 0.450^ + 0.2706 

magenta = o.o75r + 0.500^ + 0.2256 

yellow = 0.02 5r + 0.050^ + 0.5006 

The color matrix for the set as the dyes are actually present in the film, is 
therefore 



MASKING 


.90 


•45 


.27' 


.07 


•So 


•23 


•°3 


•°S 


•50, 



291 



We have rounded out the values, since a third-place accuracy is meaningless in 
photographic sensitometry. 

A color matrix such as this is a veritable fountain of information for the 
photographic technician. Let us suppose that these are the dyes to be used 
in a monopack process utilizing color development. The emulsions must then 
be balanced for contrast so that the red-sensitive emulsion will yield a cyan 
image whose red contrast is .90. The green-sensitive emulsion must be such 
that the magenta image it yields has a green contrast of 0.50. The blue- 
sensitive emulsion must yield a yellow image whose blue contrast is likewise 
0.50. Thus the color matrix of a reproduction process specifies completely the 
characteristics of the emulsions used in the monopack. 

Let us now set up the problem of separations from such a transparency. To 
compete with the one-shot camera, the transparency must yield separation 
negatives characterized by the matrix 




Consider the blue-filter separation. It will be represented by the expression 

— 0.27^ — o.2$g — 0.50ft 

since the cyan and magenta both absorb blue light indicative of densities 0,27 
and 0.23 respectively. Thus, superimposed upon the pattern of the blue 
densities (yellow image) will be the pattern of the red (cyan) and green 
(magenta) densities. In fact, the blue densities will only contribute 50 per 
cent of the total in this separation negative. To this extent we have correct 
blue rendition. 
Suppose by some method, it is possible to make a negative characterized by 

— o.27r — 0.232 + o.ooft 

and this negative is registered with the transparency. The blue filter will see 
the unmasked transparency as a picture whose density at the reference point is 
represented by 

o.27r + o.23g + 0.50ft 

In combination with the mask, the transparency will be represented by the 
expression which is the sum of the two, i.e. 

(0.27 - 0.27V + (0.23 - 0.23)2 + (0.50 + o.oo)ft 

or . .. . ... ■ „ 

o.oor + o.oog + 0.50ft 



.gom + 
.07W + 


•45^ = 
.50?* = 


.27 
•23 


get 






m = 
n = 


= .075 
= 45 





292 HISTORY OF COLOR PHOTOGRAPHY 

Thus by registering a special type of masking negative with the transparency, 
we were able to color correct it, in so far as the blue-filter separation is con- 
cerned: 

The correcting negative contains both red and green densities, in the ratio 
of 27 to 23. Were we to make a negative through the green filter, we would 
copy the densities in the inverse ratio. The red-filter separation copies the 
densities in the ratio of 13 to 1. Obviously, some combination of the two is 
needed. If the ratio of the two is m of the red to n of the green, we can formu- 
late these equations 



Solving these for m and w, we get 



Our correcting masks will be a red-filter negative represented by 

— o.o7r — o.oig — 0.006 

plus a green-filter negative represented by 

— o.2or — o.22g — 0.026 

These are made by developing the red-filter mask to a gamma of 0.08, and the 
green-filter mask to a gamma of 0.44, these being the sum of the densities 
imaged in each mask. 

The complete mask can also be made by giving the film two exposures, first 
through the red filter, then through the green. This is done to eliminate any 
possible latent image reduction by a Herschel effect. The combined masks 
register a total density of 0.08 as a result of the red exposure, and 0.44 as a 
result of the green exposure, when the development is carried to a gamma of 
0.52. The difference between them is 0.36. Were the development carried to 
a gamma of 1.00, this difference would be 0.36/0.52 = 0.70. This, then, is the 
logarithm of the relative exposures that must be given through the two filters. 
Therefore the green filter exposure is 5 times (which is the antilog of 0.70) that 
of the red. If the material being used requires an exposure of 10 seconds, with 
no filter, to yield the correct densities at a gamma of 0.52, and if the filter 
factors are 4 for the red and 6 for the green, then 40 seconds through the red, 
or 60 seconds through the green, would be the total exposures needed. But of 
the total exposure, the red constitutes only £, and the green f. Therefore the 
mask is made by first giving it a red exposure equal to ^ or 7 seconds, followed 
by a green exposure equal to 60 X f or 50 seconds. The exposed film is de- 
veloped to a gamma of 0.52. 

The mask will contain densities that will be the sum of the two masks, so 
that we can express it as 



MASKING 293 

— (0.20 + o.oj)r — (0.22 + o.oi)g — (0.00 + 0.02)6 
or 

— 0.27 — o.23g — 0.026 

This in combination with the transparency, when viewed through the blue 
filter will have for its mathematical form 

o.oor + o.oog + 0.486 

Thus we have succeeded in masking out the deficiencies due to the poor blue 
transmissions of the cyan and magenta images. 

In a similar manner the other two colors can be corrected to yield the equa- 
tions 

o.82r + o.oog + 0.006 for the red 
and 

o.oor + o.45g + 0.006 for the green 

The corrected transparency (when masked with the proper negative, and viewed 
with the proper separation filter) has the color matrix 



(0.82 0.00 o.oo\ 
0.00 0.45 0.00 ) 
0.00 0.00 0.48/ 



This is but half the job. We have succeeded in correcting for the poor spectral 
characteristics of the secondaries used in the transparency. Now we must 
balance the contrasts. 

The answer is obvious. Let us suppose we desire our separation negatives 
to have a gamma of 0.80. This means our negative should have the color matrix 



- 0.80 


0.00 


0.00 


-0.80 


0.00 


0.00 



o.oo\ 

0.00 J 

- 0.80/ 



We know now that a separation through the blue filter will yield a density of 
0.48 where the original had a density of 1.00. We must bring this value up to 
0.80. Therefore we must develop bur blue-filter separation to a gamma equal 
to 0.80/0.48 or 1.67. Similarly the green-filter separation must be developed 
to a contrast equal to 0.80/0.45 or 1.78. The red-filter separation must be 
developed to a gamma of 0.98, or practically 1.00. This means that a process 
emulsion is in order for the making of the separations. 

If these separations are converted into positive prints by the reproduction 
technique symbolized by the color matrix 



/o, 
(o, 
\o. 



,90 


0.45 


0. 


,07 


0.50 


0. 


.03 


0.05 


0. 



the reproduction will be identical with the transparency. Thus by printing the 
separations on to a monopack material of the same type as used for the 
original transparency, or one equal in function to it (having same color matrix), 



294 



HISTORY OF COLOR PHOTOGRAPHY 



identical duplication is achieved. This is of great importance in those fields 
where multiduplication is required. 

If the spectra of the secondaries are known, two of the masks can be made 
by a single exposure upon a single piece of film. Consider a set of curves such 
as those indicated in Fig. 107. These curves are not the same as those which 
give rise to color matrix A. This is done expressedly to forestall any possible 

|.0 



•4 



0.0 













C 


r-AN 




YELLOW 




MAgSNT 


* / 








/ s 













































400 



500 



600 



Cyan Magenta Yellow 



400 


.19 


O.II 


0.50 


42s 


•IS 


.15 


.60 


450 


.12 


.20 


•55 


475 


.13 


•37 


.28 


500 


.16 


.58 


.28 


525 


.23 


.68 


.19 


550 


•35 


.68 


.12 


575 


•55 


•50 


.09 


600 


.78 


.25 


.06 


625 


.88 


.15 


.05 


650 


.90 


.09 


.04 


675 


.88 


.06 


.03 


700 


.81 


.04 


.03 



700 



FIG. 107 



use of the data developed here, for an actual case. We are interested only in 
showing how the problem can be solved for the general case, not for a specific 
instance. 

The curves indicate that the cyan and magenta intersect in the region of 
wavelength 572. If we use monochromatic light of this wavelength, or a com- 
bination of Wratten filters, the center of gravity of whose transmission (in a 
very narrow band) is the wavelength, we would copy the cyan and magenta 



MASKING 295 

images in equal proportions. Our mask requires a proportion of 27 of the cyan 
to 23 of magenta, i.e. the cyan is to be copied to a slightly greater extent. This 
will be true if we shift the center of gravity slightly, say to 575 mju. At 600 mju, 
the ratio of cyan to magenta will be almost 3 to 1. At 550 m/x the ratio of cyan 
to magenta is .35 to .68 or 1 to 2. So it is seen that depending upon the wave- 
length used, we can get almost any ratio desired of cyan to magenta. 

The magenta and yellow curves intersect at about 480 m/z. Using light of 
this wavelength, we copy the yellow and magenta images in equal amounts. 
At shorter wavelengths we would copy more yellow than magenta, while at 
longer wavelengths the reverse is true. Since the yellow and cyan curves do 
not intersect in the regions of their high absorptions, we cannot make a mask 
for the green-filter separation in this manner. 

In the analysis above we used a negative density. This density is a value 
arrived at by a simple calculation. Consider the negative made of a gray scale 
whose densities vary from 0.00 to 2.00. That portion in the negative which 
corresponds to a density of 0.00 in the original, will be the blackest. Suppose 
we developed the negative to a gamma of 0.80, and that we were very careful 
to expose so that we are on the straight-line part of the curve. Suppose the 
straight-line portion starts at a density of 0.30, so that this point corresponds 
to the point in the original with a density of 2.00. Then the point correspond- 
ing to 0.00 on the original will have a density of 0.30 + 0.80 X 2.00 or 1.90. 
The point in the negative which corresponds to the density of 1.00 in the 
original, will have a density of 0.30 + 0.80 X 1.00 = 1.10. The difference be- 
tween these densities will be 1. 10 — 1.90 = — 0.80. This is the negative density 
which we used in our equations above. It is the difference in the densities in 
the negative, between a pure white and the point under consideration. If that 
point has a density of 1.00, then the actual value obtained is likewise a measure 
of the contrast to which the negative was developed. In the case of separations 
made from transparencies, where light is absorbed by all three layers, each 
layer will contribute to the total, so that the negative contrast to which a 
film was developed which has for its form 

— o.gor — o.o7g — 0.036 
would be 0.90 + 0.07 + 0.03 or 1.00. 



CHAPTER 20 
CHEMICAL TONING 



Th: 



.HE black-and-white photographic image is composed of finely divided 
silver. In such condition it is an excellent reducing agent. If the oxidation 
is carried out in the presence of a silver ion precipitant, the metallic image 
becomes converted into one composed of silver salts. These can be made to 
undergo other reactions, such as conversion into a salt of a different metal. 
In this manner it becomes possible to convert the silver into practically any 
desired insoluble salt, which may itself be colored, as are the heavy metal 
ferrocyanides, or can be converted into such, as is the case with nickel ferro- 
cyanide. Processes depending upon this type of procedure, are known as 
"chemical toning." 

By far the most popular form of toning starts with the conversion of the 
silver image into a mixture of silver and some other insoluble metal ferro- 
cyanide. The reaction between silver and ferricyanide proceeds in accordance 
with the following scheme: 

4 Ag + 4 K 3 Fe(CN) 6 -> Ag 4 Fe(CN) 6 + 3 K 4 Fe(CN) 6 

For every atom of silver, one molecule of potassium ferricyanide is needed. 
As a result, the silver is precipitated as silver ferrocyanide, which localizes 
but one of the four ferrocyanide groups formed. The other three pass into 
solution and are washed out of the film. Therefore this method is very ineffi- 
cient, since the greatest part of the desired reaction product is lost. No 
harm is done if it is desired to re-develop, since it is the silver salt which re- 
acts here, and all tne silver is used up. But if the purpose is to form the 
colored ferrocyanide salts of iron, copper, uranium, nickel, vanadium, etc., 
then only a 25 per cent efficiency is achieved. 

The early experimenters evidently recognized this fault, for efforts were 
soon made to overcome it by concocting solutions which would precipitate 
the ferrocyanide ions as fast as they are formed. This is accomplished by 
having soluble salts of the desired metal present in the solution. If it is 
desired to form nickel ferrocyanide, the solution is compounded to contain 
nickel nitrate'. In that case, the reaction would go in accordance with this 
equation: 

4 Ag + 4 K 3 Fe(CN) 6 + 6Ni(N0 3 ) 2 -* Ag 4 Fe(CN) 6 + 12KNO3 + 3Ni 2 Fe(CN) 6 

Both silver and nickel ferrocyanide are insoluble, so that all of the ferrocyanide 
is precipitated. But not all of it exists in usable form. For instance, if the 

296 



CHEMICAL TONING 297 

color itself consists of a nickel compound, then only three of the four ferro- 
cyanide molecules are used. This is much better than the first case, where 
only that part of the total is available which exists as part of the silver ferro- 
cyanide molecule. 

If an ion is present in the toning solution which precipitates silver more 
easily than the ferrocyanide, then it becomes possible to use up all the ferro- 
cyanide. The reaction then proceeds according to this schedule: 

4 Ag + 4 K 3 Fe(CN) 6 + 8Ni(Br) 2 -+ 4 AgBr + 4 Ni 2 Fe(CN) 6 + i 2 KBr 

Now all four molecules of ferrocyanide are precipitated as the nickel salt, 
which can then be converted into the color by reaction with di-methyl-glyoxime. 
We can now compare the three possibilities. In the first case only one molecule 
of ferrocyanide is available, since all the rest is removed in the wash which 
must always follow the treatment with ferricyanide. By treatment with 
nickel halides, the silver ferrocyanide becomes converted into the nickel salt 
thus: 

Ag 4 Fe(CN) 6 + 2NiBr 2 = 4 AgBr + Ni 2 Fe(CN) 6 

In the second case, three molecules of nickel ferrocyanide are formed directly. 
If the print be then further treated with nickel bromide, the fourth molecule 
is formed. In the third case, the four molecules become formed immediately. 

It is apparent that this procedure can be utilized to intensify a weak silver 
image. It is merely necessary to treat the nickel-silver ferrocyanide image 
with silver nitrate, to cause the nickel salt to become converted into the silver 
salt. Since in the second case there are three nickel ferrocyanide molecules, 
containing six nickel ions, there will be required twelve silver ions, thus 
multiplying the silver content four times. 

Not only are the heavy metal ferrocyanides insoluble, but the corresponding 
ferricyanides are also. This makes the devising of toning baths a difficult 
matter, but by the application of a few elementary principles of physical 
chemistry, their formation becomes possible. Let us consider these principles. 
A salt like NaCl can be represented as a union of a positively charged sodium 
ion, Na + , with a negatively charged chloride ion, , Cl~. This can actually 
be shown to be the case by the imposition of a strong electric field about liquid 
NaCl. It will be found that the sodium will migrate to the negative pole, 
while the chlorine will migrate to the positive pole. When salt is dissolved 
in water, this type of dissociation also takes place, and if an electric field be 
imposed upon a solution of NaCl in water, it will be found that the sodium 
will migrate to the negative pole, while the chlorine will go to the positive. 
This indicates that under the imposition of the field, the two electrically 
charged ions dissociated to form free positively and negatively charged par- 
ticles. The strength of the field required to completely dissociate the ions, 
depends upon the surroundings. Thus to dissociate molten sodium chloride 
will require very strong currents, while relatively weak ones will suffice in the 



298 HISTORY OF COLOR PHOTOGRAPHY 

case of a solution. The reason for this is that water has a dielectric constant 
equal to a value of approximately 80. 

When two charged particles, of opposite sign, are a distance d apart, the 
force of attraction between them is given by Coulomb's law: 

1 eiXe 2 



F = 



c d? 



where F is the attractive force, e\ and e^ the numerical value of the charges, 
and c, a constant, called the dielectric constant, of the medium in which the 
charges are dispersed. This law holds true not only for isolated charges in 
space, but also for ions in solution. It is therefore seen that the force which 
holds the sodium and chloride ions together in the solid state, will be much 
greater than when the salt is dissolved in water, for air has a value of 1.00 
for the dielectric content. 

That property of water, which gives rise to a value of 80 for the dielectric 
constant, acts upon all charged particles which are dispersed in it. The action 
is to weaken the bond between the ions. In many cases, this weakening may 
be sufficient to allow the ions to become entirely free from each other, especially 
under the constant impacts received from collisions due to Brownian move- 
ment. This phenomenon is called dissociation, and the extent to which this 
takes place is called the degree of dissociation. Very weak solutions of salts 
are usually considered to be dissociated 100 per cent. f Strong solutions are 
very seldom completely dissociated. Water, HOH, also dissociates very 
slightly to form hydrogen ions H + , and hydroxyl ions OH"". When a sub- 
stance in its pure form dissociates^ just as many negative as positive ions are 
formed, so that the solution remains in an electrically neutral condition. 

When a substance AB, dissociates in water, an equilibrium is set up, which 
may be expressed by the equation : 

(4+) X (B~) , + 

- — ' . -X — = a constant 
(AB) 

The value for the constant is specific for each substance, and it is a func- 
tion of temperature only. Here (A+) is the measure of the concentration of 
the ion A + , (B~) of the ion B, and (AB) of the undissociated molecule AB. 
If water is the substance considered, then (AB) or (H2O) is a constant, and 
the equation can be written: 

(H+) X (OH") « io- 14 

Since (H + ) = (OH""), this equation can be solved for (H + ), 

(H+) 2 = io" 14 , or 
(H+) - io" 7 

Water at a temperature of 25 C will dissociate to such an extent that the 
solution will have a concentration of hydrogen ions equal to 0.0000001 N. 
There will also be an equal concentration of hydroxyl ions. 



CHEMICAL TONING , 299 

Chemists have adopted a very convenient nomenclature for this value. The 
logarithm to the base 10 of this number, is minus seven. The negative of the 
•logarithm is plus seven. This value, the negative of the logarithm of the 
molar concentration of hydrogen ions is written pH, 

pH = - log (H+) = log : 



(H+) 

A neutral solution will have a pH value of 7.0. 

The most general law of solutions can now be stated. Whenever any sub- 
stance is dissolved in water, the product of the concentrations of the hydrogen 
by the hydroxyl ions must always be io~ 14 . Since the first is measured by 
pH, and the second by p(OH), and since these represent logarithms, 

pH + p(OH) = 14 

In a neutral solution, pH = p(OH), so we must have pH = 7.0. If the hydro- 
gen ions predominate, then the value of pH will be numerically lower. Acid 
solutions will have pH values that are less than 7.0. If the (OH) ions pre- 
dominate, then p(OH) will be less than 7, so that the pH, in order to maintain 
the constancy required by the equation above, will be greater than 7.0. Alka- 
line solutions will be characterized by a value for the pH that is greater than 7.0. 
Every solid substance has a definite and limited solubility in water. Let us 
center our attention upon the substances with very low solubilities, such as 
the silver halides, which have the following solubilities. 



AgCl 


Silver chloride 


8.9 X io" 5 


0.62 X io" 5 


AgBr 


Silver bromide 


8.4 X 10-* 


0.44 X io -6 


Agl 


Silver iodide 


3.0 X io~ 7 


10.3 X io~ 7 



The values in the third column represent the solubility (in grams) of the sub- 
stance in 100 cc of water. The values in the last column represent the molar 
concentrations. For ions in solution, in equilibrium with solid matter, we 
can write the following equation: 

(A+) a + (B-) b = constant, 

where the chemical formula of the substance is A aB b . In the case of AgCl, 
both a and 6 are equal to one. The value for the constant would therefore be: 

0.62 X icr 5 X 0.62 X io~ 5 or 0.37 X io~ 10 

since the molar concentrations of Ag 4 " and Cl~ are identical in a pure saturated 
solution of AgCl in water. The values for the solubility product for some 
typical insoluble silver salts are: 



AgCl 


0.37 X io" 10 


AgCNS 


0.49 X 10 12 


AgBr 


0.20 X io~ 12 


Ag 2 S 


1.6 X 10- 49 


Agl 


0.12 X io~ 15 


HgS 


4 X io- 63 



The importance of the concept of solubility product lies in this statement. 
If two solutions each containing one of the ions of a difficultly soluble salt 



300 HISTORY OF COLOR PHOTOGRAPHY 

are mixed, no precipitate will be formed unless the product of the ion concen- 
trations in the mixture is greater than the solubility product. Thus, when a 
solution of silver nitrate is added to a solution of potassium chloride, no pre- 
cipitate will take place until the concentrations of Ag + and Cl~ are sufficient 
to satisfy the equation 

(Ag+) X (CI") = 0.37 X 10- 10 

It must be carefully borne in mind that the equation above demands the 
silver to be present as a silver ion. If something is added to the solution which 
represses that ionization, a buffering action can be noted. Let us consider 
the case of silver hydroxide formed by the addition of ammonia to silver salts. 
The very first addition of ammonia causes this reaction to take place 

AgN0 3 + NH4OH -> AgOH + NH4NO3 

The solubility product of silver hydroxide is 1.52 X io~ 8 . Since silver nitrate 
is completely dissociated in a solution whose concentration is 0.1N, the molar 
concentration of silver ions would be io~ 2 . Therefore if the ammonia disso- 
ciates to produce hydroxyl ions to the extent of io~ 7 , a precipitate would form. 
Ammonia, in normal solution, dissociates to the extent of 0.4 per cent, so 
that a molar solution of ammonia would give rise to 0.004 moles of OH ions. 
The (OH) concentration would then be io~ 3 , which is considerably greater 
than the value required to satisfy the solubility product law. Therefore pre- 
cipitation would take place. But immediately a second reaction takes place, 
typified by the equation 

AgOH + 2NH 3 -> [Ag(NH 3 ) 2 ]OH 

This new substance is extremely soluble, but in solution it dissociates in a 
peculiar manner 

[Ag(NH 3 ) 2 ]OH +± CAg(NH 3 ) 2 ] + + OH~ 

No longer are there present silver ions, except to the extent that is demanded 
by a new relationship 

(Ag+) X (OH-)(NH 3 ) 2 _ 
(Ag(NH 3 ) 2 OH) 

The addition of ammonia reduces to a very small fraction the concentration 
of the silver ions present in solution. This disturbs the equilibrium condition 
between (Ag + ) and (OH~) so that the product must necessarily become less 
than io~ 8 . In order to preserve this value, some silver hydroxide will go into 
solution. If the concentration of ammonia is sufficient to react with the 
dissolved silver, it will immediately go back into solution and more silver 
hydroxide will be forced to dissolve. In this manner it becomes possible to 
dissolve completely all the silver. The solubility of silver chloride in ammonia, 
potassium cyanide, sodium sulphite, potassium iodide, hypo, etc., can all be 
explained in a similar manner. There are formed substances with very limited 
dissociation into silver ions, but which, nevertheless, are very soluble in water. 



CHEMICAL TONING 301 

When silver ions are added to a solution that contains two groups that 
form insoluble salts with the silver, the salts will be formed in the inverse 
ratio to the solubility products of the two. This comes into play, for instance, 
in the case when a silver image is oxidized by means of a solution of ferricyanide. 
Both silver ferrocyanide and ferricyanide are insoluble, but since there are 
four silver atoms in the ferro salt and only three in the ferri, the solubility 
product for the first varies with the fourth power of the silver concentration, 
and for the second it varies with the third power. Therefore the value for 
the ferro salt will be much lower than the value for the other. As the silver 
and the ferri react, silver ions are formed together with ferrocyanide ions. In 
solution, at the instant of reaction, there will be ferricyanide, ferrocyanide, 
and silver ions present, the first being present in relatively high concentrations, 
the other two in very small amounts. Equilibrium between the ions will 
quickly be established, in accordance with the relationships 

(Ag+) 4 X (Fe(CN) s 6 ) = ki, and 
(Ag+) 3 X (Fe(CN) 6 -) = k 2 

As soon as enough silver and ferrocyanide ions are formed to satisfy these 
relations, the two insoluble salts will begin to be precipitated inversely to the 
ratio of the values ki and k 2 . Thus if k 2 is io 3 times as great as ki, the silver 
ferrocyanide, corresponding to the constant ki will be precipitated io 3 times 
as rapidly as the ferricyanide. The final image will consist of 999 parts of 
ferrocyanide and one part of ferricyanide. For all practical purposes, this 
can be considered as consisting of pure ferrocyanide. 

When a silver image is treated with a mixture of potassium ferricyanide and 
chloride, the image will consist of silver chloride only. This would indicate 
that the solubility product for silver chloride, which is 0.37 X io~ 10 , is con- 
siderably lower than the value for silver ferrocyanide. The value for the 
bromide is 0.20 X io~ 12 , a value that is about sixty times less than that of the 
chloride. Therefore when both chloride and bromide ions are present in a 
solution it will be predominantly silver bromide that will be precipitated, 
with an appreciable mixture, approximately i\ mole per cent, of the chloride. 
The value for the iodide is 0.17 X io~ 15 , a thousand times less than the value 
for the bromide. 

These principles can be applied in another manner. The conversion of a 
silver image to one of silver chloride is very easily accomplished by use of 
copper salts plus sodium chloride. But it may be desirable to have an image 
of thiocyanate, or iodide, since these substances can absorb basic dyes while 
the silver chloride will not. If the silver chloride image be treated with a 
solution of potassium thiocyanate, or potassium iodide, a reaction will take 
place in this manner: As soon as the solution has saturated the film, enough 
silver chloride will dissolve to satisfy the equation 

(Ag+) X (CI") = 0.37 X io" 10 



3 o2 HISTORY OF COLOR PHOTOGRAPHY 

But since there are present iodide ions, a new equilibrium will be established, 
namely: 

(Ag+) X (I") = 0.17 X io" 15 

The less soluble silver salt of the two, the one with the lower value for the 
solubility product, will be precipitated and as long as there is AgCl left this 
reaction will continue, until finally practically all the chloride will have been 
converted into the iodide. It is not possible to convert the iodide into any 
salt whose solubility product will be greater than 0.17 X io -15 , unless there 
are added substances to the solution which form non-dissociated silver com- 
plexes with the. iodide. Under such conditions the value for (Ag*) can be 
made so low that even with iodide present in the solution, the product will be 
less than 0.17 X io~ 15 . This happens when silver salts are treated with hypo, 
cyanide, thiourea, thiocyanate, very large excesses of iodide (to the order of 
10 per cent or more), and many other salts. In these cases there are formed 
silver complexes in which ionization into silver ions becomes reduced to a 
value that is a minute fraction, one thousandth or less, of the former value. 
Then, regardless of whether this non-dissociated silver complex is soluble or 
not, the silver ions become depressed to practically the vanishing point. 

The principle just enunciated is used to a considerable extent in compound- 
ing single-solution copper and iron toners. Cupric and ferric ferricyanides 
are insoluble in water. Therefore should ferric or copper salts be added to 
ferricyanide solutions, an immediate precipitate is formed. But both these 
metals form non-dissociated salts with citrates, oxalates, etc. Therefore when 
it is desired to form single-solution toners which contain cupric or ferric salts 
together with the ferricyanide, it is essential that citrates or oxalates be present. 
The most convenient method to use with ferric salts is to utilize ferric am- 
monium citrate or oxalate. With copper salts it is necessary to have present 
from fifty to one hundred grams of potassium oxalate or citrate per liter of 
solution. Then the ferric or cupric ferricyanide will no longer precipitate. 

A similar scheme is used with iron developers. It will be recalled that 
when exposed silver halides are treated with ferrous salts, metallic silver is 
formed, the ferrous ion being simultaneously converted into ferric ions in 
accordance with the equation: 

Ag + + Fe++^ Ag° + Fe+++ 

This is a reversible reaction. When halides are present, the concentration 
of the silver ions will be considerably depressed, causing the reaction to pro- 
ceed to the left. Therefore ferrous sulphate cannot be used as a developer, 
since the tendency would be for the ferric salts formed to immediately attack 
the silver to yield silver halides again. But when potassium oxalate is added 
to ferrous sulphate in sufficient concentration, there is formed the ion 
Fe(C 2 04)2 = , present to the extent required by the equilibrium 

K 2 Fe(C 2 4 )2 *=* FeC 2 4 + K 2 C 2 4 



CHEMICAL TONING 303 

The ferrous oxalate is present in free form, only to a very mild extent, but in 
sufficient quantity to have present an appreciable amount of ferrous salts. 
The complex K 2 Fe(C 2 4 )2 is quite soluble, but very, very little dissociated, 
sufficiently low to enable the otherwise insoluble ferrous oxalate to remain in 
solution. The corresponding ferric salt K3Fe(C 2 4 ) 3 is also soluble, and but 
slightly dissociated to form ferric ions. A condition similar to the one noted 
for the use of the silver ferro- and ferricyanides, is true here. The instability 
constants in the two cases 

Fe 2 (C 2 4 ) 3 + 3K 2 C 2 4 *=* 2K 3 Fe(C 2 4 ) 3 and 
FeC 2 4 + K 2 C 2 4 ^ K 2 Fe(C 2 4 ) 2 

depend upon the third powers of the concentration of potassium ions in the 
case of ferri salts, and upon the second power in the case of the ferrous salts. 
Therefore the ionization of the ferric ion will be considerably reduced in a 
mixture of the two salts, leaving the concentration of the ferrous ions to be a 
maximum. Now when a silver image is treated with potassium ferrous 
oxalate, there will be sufficient ferrous ions formed to react with the silver 
salts. As fast as the ferric salts are formed, they unite with excess oxalate 
ions to form the complex ferric oxalate ion, that is relatively unionized. As 
fast as the ferrous salt is removed by the conversion into the ferric ion, more 
of the ferrous oxalate ion will dissociate, since equilibrium must be maintained. 
Therefore the presence of a large quantity of the very weakly dissociated 
complex, Fe(C 2 4 ) 2 5= , acts as a reservoir of ferrous ions. This is called a buffer 
action. Its greatest use is in the control of the pH of a solution. 

As was pointed out above, whenever water is present, in the water phase 
there must exist this equilibrium 

(H+) X (OH") = 10- 14 

If a mole of an acid which is completely dissociated, is dissolved in water, there 
will be present io° or 1 mole of hydrogen ions. In that case the value for 
(H + ) will be io° or 1, and the value for the (OH~) will be io~ 14 . We can write 
this in the form (cf . above) 

pH = o 
p(OH) = 14 

Consider the case where 6.2 grams of boric acid are dissolved in water. 
This is sufficient to make a tenth-molar solution. Were this acid completely 
dissociated, the pH of this solution would be approximately one. But boric 
acid in this concentration dissociates only to the extent of 0.01 per cent. 
Therefore instead of being tenth-molar with respect to hydrogen ionization, 
it is 0.01 per cent of tenth-molar, or 0.00001 molar. This would correspond 
to a pH of approximately 5.00. Despite the fact that a sufficient quantity 
of boric acid is dissolved in solution to be tenth-molar, the hydrogen ions 
are present only to the extent of 0.01 per cent of the total. Suppose now that 
to this solution is added something which uses up the acid or hydrogen ions. 



3 04 HISTORY OF COLOR PHOTOGRAPHY 

Since equilibrium must be maintained under all circumstances, as soon as 
some hydrogen ions become used up, more. boric acid will dissociate. The 
dissolved but undissociated boric acid acts as a reservoir of hydrogen ions, 
feeding these to the solution as rapidly as they are used up. This principle is 
made use of in the concoction of buffered developers, a subject which does not 
concern us at this particular moment. 

We have indicated above how the principles of dissociation, ionization, 
solubility product, and instability constant, are applied to the preparation of 
single-solution toners. It would sound well were it possible to state that 
the persons who developed the toning processes were fully familiar with these 
ideas, and that their work was the result of a careful analysis of the various 
constants involved. But that would be a gross mistatement. The solutions 
were compounded empirically. It was discovered that the addition of oxalates 
prevented the deposition of insoluble ferrous oxalate. It was natural therefore 
to apply the same technique to the preparation of other solutions that would 
form insoluble iron salts. Surely in 1840, when Herschel first discovered the 
cyanotype process (now known as blue printing), these ideas were unheard of. 
And yet he used ferric ammonium citrate in solution with potassium ferri- 
cyanide. 

This disclosure represents probably the first case in history where an image 
in color was produced. True, the color was a monochrome, blue-green, but 
it paved the way for all the others. This reaction, in this or some modified 
form, has been patented and repatented a countless number of times. As soon 
as it was generally recognized that the silver of the image could replace light, 
the reaction was again patented in any number of new variations. By 1881 
and 1882, when C. R. Woods published his notes concerning the conversion 
of the silver image into colored metallic ferrocyanides (Brit. J. Phot., Vol. 28 
(1881), p. 675, Vol. 29 (1882), p. 248) the reaction was so well known that it 
did not create any more than a passing notice. The Woods technique was to 
treat the silver image with a solution of ferricyanide. This, as was indicated 
above, formed an insoluble silver ferrocyanide, which was next converted 
into the corresponding copper, molybdenum or other heavy metal salts. 
Iodides could also be used. Mr. Woods did not propose his procedure to be 
used for the purposes of color photography. 

This was done by R. Namias (Phot. Korr., Vol. 51 (1894), p. 323). Prof. 
Namias used the light-sensitive properties of the toning solutions. A paper 
was sensitized with a saturated solution of lead lactate to which had been 
added 10 per cent of potassium ferricyanide. The lactate ion maintained the 
lead ferricyanide in solution by depressing the ionization into lead ions. After 
exposure, the ferricyanide became reduced to ferrocyanide. Since the solu- 
bility product for the ferrocyanide salts is considerably lower, as a rule, than 
the values for the corresponding ferricyanide, the presence of the lactate ions 
did not interfere with the precipitation of that salt. After exposure, the 
paper was washed clear of unreacted ferri, then treated with sodium or po- 
tassium dichromate. The lead ferrocyanide reacted with this substance to 



CHEMICAL TONING 305 

form the yellow lead chromate, probably the finest yellow pigment that could 
be used in three-color photography. The paper was washed free of dichromate, 
then sensitized with a standard blue printing solution containing ferric am- 
monium citrate and potassium ferricyanide. This was exposed under the red- 
filter negative, to yield the blue-green image. The paper was finally sensitized 
with copper lactate and ferricyanide. Here were noted some difficulties. It 
was not possible to prepare a completely clear copper ferricyanide solution 
unless alkalis were present. But this also had a solvent action upon the brick- 
red image that was formed. For this reason, it was not possible to completely 
fix the red image. 

If the paper were sensitized with the lead lactate solution used for the 
yellow image, this difficulty was overcome because then an insoluble but 
fixable image of lead ferrocyanide would result. This could' be converted 
into the brick-red copper salt by treatment with cupric acetate or sulphate. 
Another possibility, also pointed out by Prof. Namias, would be to prepare 
silver images from the three separation negatives. The one made from the 
blue-filter negative was treated with the lead lactate-potassium ferricyanide 
solution. The silver reduced the ferri to ferro, forming a mixture of the lead 
and silver salts, as indicated above. The red-filter positive was treated with 
the usual cyanotype solution, and this gave a blue-green image directly. 
The image from the green-filter negative was converted first into silver ferro- 
cyanide, then into copper. 

Many variations existed for the formation of the yellow and magenta images, 
but only one general reaction was used to form the blue-green impression. 
This was to convert the silver into ferric-ferrocyanide, or Prussian blue. 
The reaction was first disclosed by that early photographic scientist, Sir John 
Herschel, in 1840. He discovered that a mixture of ferric ammonium citrate 
and potassium ferricyanide was sensitive to light, especially in the presence 
of organic matter, such as paper. A paper, sensitized with this solution, and 
exposed under a negative, gave a blue-green image of Prussian blue wherever 
the light reacted with the chemicals. The unreacted salts could be washed 
out, leaving clear white areas. The Herschel idea was utilized by J. Lewisohn 
(U.S.P. 1071559 and 1126495; Eng. P. 2474/15) for all three images. Thus 
each component is first printed upon a cyanotype emulsion. Three blue- 
green iron ferrocyanide images result. The print made from the red-filter 
negative is left undisturbed. The blue-filter positive is dyed with aurantia, 
then treated with silver nitrate. This decomposes the cyanotype image, 
and fixes the dye in the image-bearing portions. The green-filter positive is 
dyed with eosine red and then treated with silver nitrate. 

It was soon discovered that the action of light could be replaced by a silver 
image. Generally it is sufficient to treat the silver image with 

Ferric ammonium citrate 10 parts 

Potassium ferricyanide 10 parts 

Acetic acid 10 parts 

Water to 1000 parts 



306 HISTORY OF COLOR PHOTOGRAPHY 

As pointed out above, this solution would utilize only three-fourths of the 
ferrocyanide formed. The remaining 25 per cent is present as the opaque 
colorless silver salt, which should be removed in order to have brilliant images. 
But it can also be used to change the tones from a blue-green to blue or green. 
For instance, if the blue-toned image is treated with vanadium chloride, the 
remaining silver ferrocyanide becomes converted into the yellow vanadium 
ferrocyanide. This represses the blue transmission, yielding a green image. 
R. Namias (Jahrbuch (1903), p. 158), utilized a mixture of iron and vanadium 
chlorides to obtain a green tone. After converting the image into silver ferro- 
cyanide it was treated with: 



Ferric chloride 


4.8 parts 


Vanadium chloride 


4 parts 


Ammonium chloride 


10 parts 


Hydrochloric acid, cone. 


10 parts 


Water to 


1000 parts 



The silver chloride that is formed can be removed by treatment with hypo. 

A great deal of fuss seems to have been made with regard to the presence of 
halide in the toning solution. In the absence of such salts, part of the ferro- 
cyanide is prevented from functioning, since it becomes converted to the 
translucent but ineffective silver salt. The opacity of silver chloride or 
bromide cannot be very much greater than that of the ferrocyanide, but Prof. 
E. J. Wall was sufficiently impressed by it to concoct a solution for green 
tones, that contained no halides {Phot. /., Vol. 68 (1921), p. 96; Am. 
Phot., Vol. 16 (1922), p. 396). A solution of vanadium oxalate was first 
prepared. To this end 100 grams of ammonium metavanadate is treated with 
460 grams of oxalic acid, and 500 cc of water. Upon heating, the mixture 
first forms a paste which finally becomes very fluid, the color changing in 
the meantime from white to orange-red, and finally to a gray-green. Water 
is added until solution is achieved, at which time the color changes to-a brilliant 
blue. If the volume is then made up to 1477 cc, there will be obtained a 20 per 
cent solution of vanadium oxalate, containing a slight excess of oxalic acid. 
If the sulphate is desired the oxalic acid is replaced by 197 cc of sulphuric acid, 
this being added to the ammonium metavanadate dispersed in 200 cc of water. 

Sufficient heat will be generated by the addition of sulphuric acid to the 
water, to cause solution to take place, but more heat should be applied if the 
color is not the brilliant blue. The toning bath is compounded as follows: 

20% Vanadium solution 50 parts 

Oxalic acid, saturated solution 50 parts 

Ammonia alum, saturated solution 50 parts 

Ferric oxalate solution as desired, 

Glycerine 50 parts 

Potassium ferricyanide, 10% 10 parts 

Water to 1000 parts 



CHEMICAL TONING 307 

The amount of ferric oxalate to be added is determined by the tone desired. 
Greener tones are obtained with less ferric salts, bluer tones with more. 

It is to be noted that three insoluble ferrocyanide salts can be formed here — 
silver, iron, and vanadium. The exact amount of each will be determined by 
the relative insolubility products, at least as far as the iron and vanadium 
salts are concerned. Since this value is not a true constant, but varies with 
the temperature, no exact relationship, except under controlled condition, will 
be possible. This is the great danger that is ever present when it is desired to 
modify the tone of a ferrocyanide image by the formation of mixed ferrocyanide 
salts. For the people who are interested in the making of two-color prints, 
the iron-vanadium toned image can be highly recommended for the green 
image. 

For three-color work, the pure vanadium image can serve for the yellow 
impression, although it is somewhat too far on the orange side. It was R. Na- 
mias (Jahrbuch (1901), p. 171) who first proposed vanadium for toning. His 
technique was to oxidize the silver to ferrocyanide, then convert that into 
the vanadium salt. In a later article, he suggested that the tone could be 
made greener by treatment with a citric acid solution of ferric alum (Rev. 
franq. Phot., Vol. 5 (1924), p. 76). In this case the iron replaces some of the 
vanadium salt, so that the action could be stopped at any stage from the 
yellow of the pure vanadium to the cyan of the pure iron ferrocyanide. 

Besides the lead chroma te or vanadium ferrocyanide images, other types, 
such as an image composed of silver and mercury iodides, could be used for the 
yellow. A. Hamburger, inventor of the Polychromide process, claimed to 
be the first to utilize this. (Brit. J. Phot., Vol. 70 (1923), p. 183; Eng. P. 
20880/11; U.S.P. 1059867; Ger. P. 329273), but Wall, in his monumental 
"History of Three-Color Photography," (p. 416 ref. No. 36) points out that 
C. R. Woods had suggested a similar idea in his 1882 paper, and that G. Brunei, 
in his book "La Photographie en Couleurs, ,, published in 1891 (p. 108) sug- 
gested that a lead-ferrocyanide image could be toned yellow by treatment 
with a potassium-mercuric-iodide complex. But even before this, C. Glissenti 
(Phot. Times, Vol. 17 (1887) p. 68) proposed to convert a cuprous iodide 
image into a yellow by toning with mercury iodide complexes. The Ham- 
burger technique converted a silver bromide or lead ferrocyanide image into 
a yellow by treatment with the following bath: 



Mercuric chloride 


24 parts 


Potassium iodide 


48 parts 


Water to 


1000 parts 



The print remains in this solution until the desired tone is obtained. The 
action can be stopped by an acid bath. H. E. Rendall also used an iodide 
image for the yellow (Brit. J. Phot., Vol. 70 (1923), p. 167). The silver was 
bleached by treatment with a 2.3 per cent solution of dichromate containing 
1.25 per cent hydrochloric acid. After washing it was treated with a solution 



3 o8 HISTORY OF COLOR PHOTOGRAPHY 

prepared by adding mercuric chloride to i2§ per cent potassium iodide until 
precipitation started. E. C. G. Caille (Eng. P. 15050/08) converted a silver 
iodide image into a yellow, by bathing with tartar emetic, then with mercuric 
chloride. J. F. Shepherd (Eng. P. 175003) in his Triadochrome process treated 
the silver image with mercuric iodide, ferricyanide, and bromide. The cyan 
was the usual cyanotype image, but he resorted to carbon to prepare the 
magenta. E. A. Lage (U.S.P. 1623123; Eng. P. 188692), used lead chromate 
for the yellow. This was formed by toning a bromide print made from the 
blue-filter negative. The paper was then sensitized with a cyanotype solution, 
and exposed under the red-filter negative. The magenta was formed by sen- 
sitizing the bromide with dichromate, and after exposure and removal of the 
unreacted dichromate salts, it was inked with a greasy magenta ink which 
took only in the exposed regions. M. S. Procoudin-Gorsky (Eng. P. 168100) 
converted the blue-filter image into a yellow by treating with potassium 
ferricyanide, potassium iodide, and ammonia. In the same disclosure he 
speaks of converting the yellow-sensation positive into a blue. This would 
indicate a very poor knowledge of the fundamental principles of color separa- 
tion. It is the red, and not the yellow impression which must be so colored. 
A pure yellow impression has no place in color reproduction, since it would 
be equivalent to the combined red and green sensations. 

A very close second to lead chromate for the yellow impression, lies in the 
use of cadmium sulphide. This is disclosed by Snyder and Rimbach in several 
patents (U.S.P. 2171609; Eng. P. 469133). Two stock solutions are used: 



Solution A: 




Cadmium nitrate 


40 parts 


Water to 


200 parts 


Solution B: 




Potassium citrate 


300 parts 


Citric acid 


30 parts 


Formalin 


100 parts 


Water to 


800 parts 



Solution A is added to B under rapid stirring, which is continued until a clear 
solution results. The bleaching bath is compounded by adding 20 parts of a 
15 per cent solution of potassium ferricyanide to every 100 parts of the above. 
The print is bleached in this to completion. Since there are no halide ions 
present, some of the ferrocyanide will exist as the silver salt. To completely 
convert this into the corresponding cadmium salt, the print is removed from 
the solution and replaced there after the addition of 10 parts of 7 per cent hypo 
for every 50 parts of toning solution. The weak hypo slowly dissolves out the 
silver ferrocyanide, which is immediately precipitated as the cadmium salt. 
A very delicate balance must be preserved between the rate of solution of the 
silver salt and the rate at which precipitation of the cadmium salt takes place, 
otherwise there is danger that not only will loss of density result, but also 



CHEMICAL TONING 309 

loss of definition. This will take place if the silver is removed much faster 
than precipitation can take place. 

After a thorough wash, the cadmium ferrocyanide is converted into the 
sulphide by treating with: 



Sodium sulphide 


5 parts 


Sodium bicarbonate 


7 parts 


Water to 


100 parts 



A yellow image of cadmium sulphide is formed. Any yellow stains in non- 
image portions that are not caused by deposition of cadmium sulphide, can 
be removed by treatment with a one and one-half per cent solution of hydro- 
chloric acid. This is the yellow that is used in the Chromatone process, which 
is based upon the disclosures of Snyder and Rimbach. More will be said 
concerning this, below. 

The magenta or red image can be prepared directly by toning with uranium. 
This substance was first used in 1859, when J. C. Burnett (Eder Handbuch, 
IV, 4, (1929), p. 159) prepared prints by exposing a sheet of paper sensitized 
with uranium salts, under a negative. The light reduced the uranium. The 
image was developed by treating this with ferricyanide, silver nitrate, gold 
thiocyanate, etc. When potassium ferricyanide was used as the developer, a 
reddish brown image resulted, that was much more suitable for two-color 
photography than for three. 

A comprehensive study of uranium toning was published by Dr. Sedlaczek 
(Phot. Ind. y (1924), p. 234; Am. Phot., (1925), p. 8). The tones obtained 
vary from a dark brown to a bright brick red. The brown tones are obtained 
from a solution containing a minimum of oxalate, 



Uranyl nitrate 


5 parts 


Ammonium oxalate 


10 parts 


Hydrochloric acid, 10% 


10 parts 


Potassium ferricyanide 


2 parts 


Water to 


1000 parts 



By increasing the oxalate content to a maximum, such as the use of 100 parts 
of saturated oxalic acid, the tone becomes much brighter. He recommended 
especially the following formula : 



Uranyl nitrate 


5 parts 


Potassium citrate 


5 parts 


Potassium ferricyanide 


2 parts 


Ammonia alum 


10 parts 


Hydrochloric acid, cone. 


0.3 part 


Water to 


1000 parts 



Since hydrochloric acid is used, the silver is converted into silver chloride. 
Hence fixation should follow the treatment. If cobalt salts be added to the 
hypo, colder tones are obtained, possibly by the formation of some cobaltous 



310 HISTORY OF COLOR PHOTOGRAPHY 

ferrocyanide, which is gray-green in color. A typical bath would be (Neblette 
" Photography, Principles and Practice," p. 448) 



Hypo 


5 parts 


Potassium citrate 


1 part 


Cobalt nitrate 


1 part 


Water to 


1000 parts 



Marines and Godowsky in their earlier patents, (U.S.P. 1516824 and 1538996; 
Fr. P. 587395) utilized toning to separate the two images in their monopack. 
This, it should be recalled, consisted of a plate with two emulsions coated 
one on top of the other. The bottom emulsion was panchromatic, while the 
upper one was orthochromatic. After development and fixation, the two 
images were converted into silver ferrocyanide. Then the upper one only 
was redeveloped. This left the lower image in the form of silver ferrocyanide, 
while the upper one consisted of metallic silver. Treatment with ferrous salts 
converted the lower image to a cyanotype, after which the upper image could 
be converted into a red by means of copper, uranium, or vanadium toning 
followed by a bath with a basic dye. The chemistry here is a little weak, 
since silver ferrocyanide when treated with ferrous salts would not produce 
an iron blue. The ferric salt must be used to accomplish this result, and 
ferric salts would be dangerous, since they would reduce the intensity of the 
silver image in the upper layer. Burwell of the Eastman Kodak Company, 
also used uranium to form one image in a monopack (U.S.P. 1966330). 
The exposed plate was developed with a dye-coupling developer which 
yielded a cyan dye plus a silver image in the two emulsion layers. 

3:5 di-brom-ortho-cresol 2.0 parts 

Alcohol 20 parts 

Sodium carbonate 7^ parts 

Potassium bromide 0.3 part 

Di-ethyl-paraphenylenediamine-hydrochloride 2 parts 

Water to 250 parts 

The dye in the upper emulsion layer is destroyed and the silver image 
simultaneously toned a red by treatment with: 



Uranyl nitrate 


5 parts 


Potassium oxalate 


5 parts 


Potassium ferricyanide 


2 parts 


Ammonium alum 


12 parts 


Hydrochloric acid, 10% 


10 parts 


Water to 


2000 parts 



This is practically identical to the uranium toner disclosed by Sedlaczek. A 
uranium toning solution that was supposed to yield a true magenta tone, was 
disclosed by J. N. Thompson and Friese-Greene (Eng. P. 160540). 



CHEMICAL TONING 311 

Potassium ferricyanide 12.5 parts 

Uranyl nitrate 12.5 parts 

Rose bengal 0.013 P art 

Naphthol yellow 0.013 P art 

Iodine 0.013 P ar t 

Acetic acid . 12.5 parts 

Water to 1000 parts 

It is quite difficult to understand why a yellow dye is needed in the preparation 
of a pink. 

Copper salts could also be used for the preparation of the magenta or red 
image. As is true with uranium, it is much more suitable for use in two-color 
work than in three. Its use can be traced as far back as 1876, when Eder 
and Toch wrote a paper about it in the Photo graphische Korrespondence. 
Namias (cf . above) suggested the use of copper lactate, but he was unable to 
prepare a clear solution. In 1900, Eder (Phot. Korr., Vol. 36 (1900), p. 537), 
suggested that clear solutions could be had by use of ammonium carbonate, 
but the copper ferrocyanide image was also soluble in this, so that loss of 
image resulted. Ferguson, (Phot. /., Vol. 25 (1900), p. 133) succeeded, by 
having present potassium oxalate. The British Journal Almanac gave the 
following technique, which used two stock solutions. 



Solution A: 




Copper sulphate 


6.25 parts 


Potassium citrate 


25 parts 


Water to 


1000 parts 


Solution B: 




Potassium ferricyanide 


5.2 parts 


Potassium citrate 


25 parts 


Water to 


1000 parts 



Just before use, mix equal parts of A and B. Sedlaczek also examined copper 
toning with the rigor he had applied to uranium, and he disclosed a number 
of solutions which gave tones varying from red to violet-brown. Since both 
copper and uranium toned images are so poor for the purposes of color pho- 
tography, they need not be discussed at this point. A further discussion will 
be had in the chapter dealing with dye toning. 

The convenience of the direct toning processes made it highly desirable to 
find a suitable method by which to produce a magenta image. T. T. Baker 
applied the finding of Feigl (Zeit. Anal. Chem., Vol. 74 (1928), p. 382) to the 
solution of this problem (Eng. P. 444773). Feigl had found that a magenta 
precipitate was formed when an ammoniacal solution of para-di-methyl- 
amino-benzal-rhodanine was added to silver salts. Baker proposed first of all 
to bleach the silver image with copper chloride, forming an image of cuprous 
and silver chlorides. These are soluble in ammonia, forming cuprammonium 
and silver-ammonium complexes. Hence when treated with an ammoniacal 
solution of the rhodanine compound, a magenta precipitate will replace the 



312 HISTORY OF COLOR PHOTOGRAPHY 

metal salts. The rhodanine compound, although quite expensive, can be 
obtained from the Organic Chemical Division of the Eastman Kodak Company. 
Karl Schinzel proposed another attack upon the problem of toning directly. 
Just as the rhodanines can form colored insoluble salts with many heavy 
metals, so can the mercaptans (Eng. P. 500716). These are the thio-alcohols, 
and contain the — SH group. The hydrogen attached to the sulphur is quite 
acidic and can be replaced by many of the heavy metals, in which case colored 
insoluble products are formed. The silver is first converted into silver chloride 
then treated with alkaline solutions of the mercaptans. Most of the colors 
produced are reds and yellows, but it is claimed that cyans and magentas 
could also be obtained. The substances can be prepared by means of the well- 
known diazo reaction, where a diazonium salt is added to a boiling solution 
of the cuprous salt of the desired acid ion. Or some other salt can be used, 
in which case powdered copper should be added. The acid ion replaces the 
diazonium group in the compound. In this case it is suggested that a xantho- 
genate be used. This is a molecular mixture of carbon disulphide and an 
alkali sulphide. After the addition, the reaction product is subjected to a 
saponification, which yields the free thio-alcohol or mercaptan. A yellow 
image is obtained by using the naphthalide of thio-glycollic acid, while thio 
carbine yields a red. It should be mentioned that the mercaptans are very 
undesirable substances with which to work. Not only are they very poisonous, 
but they have a very disagreeable odor. The well-known odor of the skunk 
is due to a derivative of mercaptan. 

The most successful solution for the preparation of the magenta image lies 
in the use of di-methyl-glyoxime. This was devised independently by Anton 
Jasmatzi (Eng. P. 402619), F. Lejeune (U.S.P. 1963707) and Snyder and 
Rimbach (U.S.P. 2100224 and 2171609; Eng. P. 469133). The Snyder and 
Rimbach disclosures are incorporated in the Chromatone process for three- 
color prints by toning, a technique that was marketed in the United States 
by the Defender Photo Supply Co. Colin Butement (Brit. J. Phot., Vol. 84 
(1937), p. 113) gave rather simple directions for the formation of the magenta 
image using this technique. The image must first be converted into an in- 
soluble nickel salt, since the magenta compound is nickel di-methyl-glyoxime. 
This is accomplished by treating the silver image with the following bath: 

Nickel nitrate 5 parts 

Potassium citrate 15 parts 

Potassium ferricyanide 2 parts 

Nitric acid, cone. J part 

Water to 150 parts 

The time of treatment is five minutes. It should be noted that a large quantity 
of potassium citrate is present in the solution to prevent the precipitation of 
nickel ferricyanide. This is in accordance to the principles discussed in the 
first part of this chapter. After a ten-minute wash, the nickel is converted 
into the magenta compound by the action of alkaline di-methyl-glyoxime, 



CHEMICAL TONING 313 

Saturated solution of di-methyl-glyoxime 5 parts 

Sodium hydroxide, 0.4% 5 parts 

Water to 50 parts 

The print is left in this solution until no further reaction takes place, after 
which it is washed in water for two minutes, fixed in 5 per cent hypo for a 
like period of time, and washed in water again. The acid treatment is necessary 
since this image may come in contact with the blue-toned image, and partially 
destroy it. 

Snyder and Rimbach (cf . above) use an entirely different chemistry. They 
point out in their specifications that in order to successfully convert the silver 
of an image into a different salt of ferrocyanide, it is necessary to have in the 
solution an acid which is stronger than hydroferricyanic acid. This means 
that equimolar solutions of the two should give a lower pH for the added 
acid than for hydroferricyanic acid. There must also be present a salt which 
will prevent the precipitation of the ferricyanide, but not the ferrocyanide. 
Suitable acids are oxalic, tartaric, acetic, nitric, etc. Suitable salts are the 
citrates, malates, oxalates, tartrates, benzoates, succinates, etc., all of which 
must be present in relatively high concentrations. Since it is quite easy to 
replace the nickel in nickel ferrocyanide with ferric ions, to form a cyan image, 
and to replace the ferrocyanide ion by means of di-methyl-glyoxime to form 
the magenta, it is possible to use nickel ferrocyanide for both the magenta 
and cyan impressions. The silver images are treated with: 



Nickel chloride 


50 parts 


Potassium citrate 


300 parts 


Citric acid 


30 parts 


Formaldehyde 


100 parts 


Water to 


1000 parts 



To every 100 parts of this solution, there is added just before use, 20 parts 
of a 15 per cent solution of potassium ferricyanide. Because this solution 
will not keep, it is best to mix just enough for immediate use, and discard 
immediately afterward. Because chloride ions are present in this solution, 
the silver will be wholly converted into silver chloride, leaving all the ferro- 
cyanide free to unite with the nickel. 

To convert the nickel ferrocyanide into the rose-colored nickel-di-methyl- 
glyoxime salt, the bleached print, after a thorough wash, is treated with 
a di-methyl-glyoxime solution made alkaline, since nickel ferrocyanide is 
somewhat soluble in caustic. Instead of ammonia, Snyder and Rimbach used 
mono- and tri-ethanolamines. These are ammonia-like substances which 
are not volatile, hence can be stored. 



D i-methyl-glyoxirne 


10 parts 


Tri-ethanolamine 


150 parts 


Mono-ethanolamine 


50 parts 


Water to 


I pop parts 



314 HISTORY OF COLOR PHOTOGRAPHY 

The mixture of the ethanolamines and di-methyl-glyoxime is heated to 100 C, 

and kept at that temperature for 15 minutes. A red-brown solution results, 

which is then diluted with water to form 1000 parts of final solution. 

Di-methyl-glyoxime is the simplest substance in a family characterized by 

the grouping 

— C=N— OH 

I 
_C=N— OH 

These are the dioximes of the ortho diketones. This grouping appears to be 
specific for nickel and palladium. The reaction between nickel and dimethyl 
glyoxime was first discovered by Tschugaeff (Z. anorg. Chew., Vol. 46 (1905), 
p. 144; cf. also "The Application of the Dioximes to Analytical Chemistry," 
H. Diehl, published by the G. Frederick Smith Chemical Co.), and immediately 
applied by him as a test for nickel. It was soon developed that all substances 
containing the ortho dioxime grouping gave a red precipitate with nickel 
and a yellow one with palladium. Other metals also gave precipitates but not 
with such characteristic colors. 

The statement that all compounds with the dioxime group yield the in- 
soluble metal salts, must be somewhat qualified. Three types of compounds 
could be formed, depending upon the special relationship of the two hydroxyl 
groups, thus: 

— C— C— — C C— — C C 

II II II II II II 

HO— N N— OH N N N N 

\ \ \ / 

OH OH OH HO 

Anti form Amphi form Syn form 

It is only the anti form of the dioxime which gives the red precipitate with 
nickel salts. Each nickel atom unites with two dioxime molecules to give a 
compound, with the structure: 

R R 

\ / 



II 


II 


0<-N 


N— 


• \ 


/ \ 


: Ni 


\ / 


\ / 


0— N 


N->0 


II 


II 


c — 


— C 


/ 


\ 


R 


R 



It is very interesting to note that in this structure nickel is quadrivalent, con- 
sisting of two covalences and two secondary valences. 



CHEMICAL TONING 315 

The color of the precipitate is a red with enough blue transmission to make 
its use as the magenta in a three-color process just passable, but not preferable. 
The fact that the R in the formula above can be made different from a methyl 
group, gives rise to the possibility that the red may be somewhat bluer with 
other substitutions. This possibility should not be overlooked by the techni- 
cian. Also the reaction with other metal salts is very interesting. Palladium 
gives a precipitate which has been described as being lemon yellow. This 
would seem to be ideally suited for the minus-blue impression. The disad- 
vantage of the expense of using palladium would be offset by the simplicity 
of operation, and by the fact that no poisonous or disagreeable substances 
are involved. 

The enterprising chemist can investigate all compounds which contain two 
hydroxyl, amino, oxime, and other groups with a labile hydrogen, situated on 
adjacent atoms, for such groups always yield the possibility of a chelate ring 
formation. The chemistry of the chelate rings has been discussed rather fully 
by Dr. H. Diehl in an article which appeared in Chemical Reviews (Vol. 21 
( I 937)> P- 39)- It is interesting to note that the blue-green pigment recently 
hailed as the answer to the printing-ink problem, is the copper salt of a complex 
containing a chelate ring, and composed of phthalic acid units. It is known, 
for that reason, as a phthalocyanine. 

Gold was suggested for the magenta image by W. Reichel (Eng. P. 6356/03; 
Ger. P. 163326). The separation negatives were printed upon stripping paper 
and then toned. Vanadium ferrocyanide, or lead chromate served to form 
the yellow image. Iron ferrocyanide formed the cyan. No novelty can be 
found up to this point. But to form the magenta, the print is toned with gold 
sulpho-cyanate in conjunction with sodium iodide and potassium carbonate. 

Complete systems for the preparation of color prints by chemical toning 
methods have been disclosed time and time again, with but little variety in 
the technique involved. A prolific patentee in this respect was F. E. Ives 
(U.S.P. 1170540, 1188939, 1225246, 1278667, 1278668, 1376940, 1499930, 
1538816 and 1695284; Eng. P. 9954/15, 113618 and 119854). In one type 
of disclosure a yellow-dyed emulsion was used. One separation was printed 
through the base, forming an image that did not penetrate deeply into the 
depth of the emulsion. This was toned cyan by treatment with: 



Oxalic acid 


4.25 parts 


Potassium f erricyanide 


1.75 parts 


Sodium chloride 


4.25 parts 


Potassium bromide 


0.35 part 


Ferric chloride solution, U.S.P. 


1.0 part 


Water to 


1000 parts 



After this conversion, the film was resensitized by treatment with bromides, 
and it was exposed again, this time through the emulsion side. Development 
was accomplished with an acid amidol solution, which would not affect the 



316 HISTORY OF COLOR PHOTOGRAPHY 

blue-toned image already present. The second image was toned red with 
copper or uranium. 

Instead of printing the second image by means of resensitized silver bromide, 
Mr. Ives suggested that the print be sensitized the second time with dichromate. 
After exposure and removal of the unreacted dichromates, the film could be 
colored with dyes that stain only soft or only hard gelatin. Such dyes were 
well-known, being sold by the I.G. under the trade name of "pinatype." 
Fast red would stain only those portions that received light. A procedure 
not very different from this was disclosed by S. J. Cox (Eng. P. 15648/14). 
One record was printed and developed with a non-staining developer such as 
ferrous oxalate or amidol. This image was toned blue. After sensitization 
with dichromate, the print was exposed to a positive of the other separation, 
forming non-staining portions wherever the light penetrated. A positive image 
could then be obtained by staining with a pinatype dye. If a third color 
is desired, the pinatype dye is fixed by the action of copper sulphate, the print 
coated with sensitized gelatin, and the process repeated. Several years prior 
to this, T. A. Mills disclosed a two-color process which called for recoating the 
film with a silver chloride emulsion (U.S.P. 1172621; Eng. P. 28081/n; 
Ger. P. 275683). The film, after exposure, development, and toning with 
iron, was recoated with a printing-out emulsion. This, after exposure, was 
toned red with a mixture of vanadium and copper or uranium. 

In 1907 F. W. Donisthorpe utilized a uranium, iron, lead, or vanadium- toned 
image for the formation of a resist against the staining of a gelatin film by 
certain acid dyes (Eng. P. 13874/07 and 158021; U.S.P. 923030 and 1517200). 
The solutions recommended were the following : 



Uranium nitrate 


1 part 


Potassium ferricyanide 


1 part 


Water 


, 45 Parts 


solution for ten minutes, 


the negativi 


Ferric chloride 


1 part 


Glycerin 


1 part 


Water 


50 parts 


however, was: 




Vanadium chloride 


2 parts 


Potassium ferricyanide 


2 parts 


Ferric chloride 


1 part 


Glycerin 


1 part 


Ferric oxalate 


1 part 



Saturated oxalic acid 100 parts 

Water 900 parts 

After toning, it was found that the gelatin about the silver image resisted the 
action of the pinatype dyes. In 1913 (Eng. P. 7368/13; U.S.P. 1193879; 



CHEMICAL TONING 317 

Ger. P. 329509) he discovered that these solutions could serve to tone positive 
silver images, for the preparation of color prints. Simultaneously and probably 
with complete ignorance of this work, W. F. Fox disclosed practically the 
identical procedure (U.S.P. 1166121, 1166122, 1166123, 1187241, 1187422, 
1187423, 1207527 and 1256675; Eng. P. 552/14, 8728/14 and 3666/15). The 
full routine of both the Donisthorpe and Fox techniques, is as follows: A 
weak positive of one record is registered with the negative of the other. This 
is probably the first application of masking to continuous-tone color reproduc- 
tion. From such a masked negative, a print is made. Let us suppose that it 
is the red negative that is used for printing. The positive is toned green by 
treatment with one of the solutions noted above. This will harden the gelatin 
immediately surrounding the toned image. The film is then dyed generally 
with a pinatype dye, complementary in color to the toned image. The dye 
will take only in the non-image portions of the film. If there is registered with 
this another film that contains a silver record of the other separation, the dyed 
portion of the film will be modulated by this added silver image. This techni- 
que was improved at a much later date, by Emil Wolfl-Heide (Eng. P. 340278). 
Mr. Wolff-Heide surface-sensitized a color-blind negative emulsion by treat- 
ment with colloidal sensitizing agents (cf. chapter on Monopacks). The 
image produced in the upper stratum of the emulsion was toned red with 
a uranium toner. This not only converted the upper image into a red, but it 
also hardened the gelatin in the image portions, so that they would no longer 
absorb pinatype blue. In this manner a negative was produced in which 
densities in the upper emulsion layer modulated the green rays, while the black- 
and-white silver densities in the lower layer modulated the dye that gave a 
general stain to the rest of the gelatin. In the English patent Mr. WolfT-Heide 
proposed to place the dyed and toned negative in contact with the emulsion 
side of a positive. If this were exposed through the negative by means of blue 
light, the uranium-toned image would modulate the light, so that in the 
regions corresponding to the red-toned image there would be produced a 
latent image in the positive. At the same time the pinatype dye would trans- 
fer to the positive. The film was developed, and the developed silver image 
toned red. 

During a small part of 1931, the writer acted as consultant for the WolfT- 
Heide Photochemical Company. By this time the entire processing was 
changed. The negative was prepared, as previously, by surface sensitization 
of a color-blind emulsion. This gave a red-sensitive stratum in the upper half 
of the layer, and blue in the lower half. After exposure, the film was developed, 
giving two images in the two emulsion strata. The top image was toned red 
with uranium, then the rest of the film was dyed blue with pinatype blue D. 
The colored negative was contact-printed upon another surface-sensitized film. 
The red rays, modulated by the silver image plus pinatype dye, would yield 
a latent image in the upper stratum of the positive film. The blue rays, 
modulated by the red-toned uranium image, gave a latent image in the bottom 



318 HISTORY OF COLOR PHOTOGRAPHY 

portion. The processing technique was identical with the one used for making 
the negative. In the final print, therefore, there was a red-toned image that 
was supposed to represent a copy of the bottom or blue density image in 
the negative, while the blue stain modulated by the silver image was a record 
of the red tones in the colored negative. It is, of course, very obvious that the 
silver image lying in the bottom portion of the negative will modulate both 
red and blue light passing through it, so that upon copying it will be impossible 
to obtain a correct rendition of the red densities lying in the uranium-toned 
portion of the image. To obtain a true copy of these densities, it is necessary 
to print with a light to which the other record is completely transparent. 
Since it remains as a black-and-white silver image, that is impossible. 

The Ives technique of printing one image, toning it, then resensitizijig the 
film with dichromate, was patented again by C. F. Jones (Eng. P. 165380; 
Ger. P. 349944), and in a slightly varied form, by M. S." Procoudin-Gorsky 
(U.S.P. 1435283; Eng. P. 135161). L. Dufay, of Dufaycolor fame, started 
with a thin base which had an emulsion coated on each side (Eng. P. 197912). 
It was thus possible to print two images, one on each side, and tone them 
separately. Upon completion, the duplitized film, as such double-coated film 
is called, could be mounted on a paper base. The application of this type 
of film to motion pictures was made by Hernandez Mejia (U.S.P. 1562828). 
W. V. D. Kelley (U.S.P. 1753379) printed an image on one side of duplitized 
film, toned it a blue, then converted the silver to silver bromide and printed 
two more images, one on each side. These were developed with amidol, then 
toned yellow and red respectively. Previously Mr. Kelley had patented other 
modifications of the Ives's technique (U.S.P. 1561168 and 1787201; Eng. P. 
228887). These related to two-color prints in which iron was used to tone one 
image and uranium to tone the other. The procedure was patented once 
more by the Colour Film Ltd., in the late nineteen-thirties (Eng. P. 400251). 
The treatment of a silver halide emulsion with acid ferricyanide solutions, 
essential in toning, destroys a good deal of the sensitivity of the film. This 
can be restored by treatment with sulphite, bisulphite, hydrazine, or hydroxy la- 
mine (Eng. P. 404856 to I.G.). Kelley, Ives, and the Eastman Kodak Co., 
in many of their patents, noted this fact. J. B. Harris Jr., printed an image 
through the base, converted the silver to copper ferrocyanide, then printed 
an image on the surface of the emulsion, and toned it with iron (U.S.P. 
1825863). In a later disclosure (U.S.P. 1848717) the bottom image was 
converted into a copper-mordant image. The sensitivity of the emulsion was 
restored by treatment with a solution containing NH 4 Br and dichromate. 
This destroys the latent image left in the emulsion. The second image was 
developed, then the first image converted into a blue by treatment with basic 
dyes, and the second toned a red. 

For a time, motion pictures in two colors were made by Multicolor and its 
successor, Cinecolor. These utilized duplitized film. The two separations 
were exposed in a camera simultaneously, in the form of a bipack. One separa- 



CHEMICAL TONING 319 

tion was then printed on each side of the duplitized positive stock, which 
was dyed yellow to prevent any printing light from penetrating from one 
emulsion to the other. After development and fixation the film was floated 
on top of a solution (so that only one side was treated) which contained: 

Ferric ammonium oxalate 9 parts 

Potassium ferricyanide 4 parts 

Ammonium chloride 8 parts 

Hydrochloric acid 4 parts 

Water to ' 1000 parts 

At the end of five minutes, all the silver in the one layer became converted 
into a blue-toned image. The film was then washed, and treated with a 
uranium toner, such as: 



Uranyl nitrate 


3 parts 


Potassium oxalate 


8 parts 


Potassium ferricyanide 


2\ parts 


Hydrochloric acid (cone.) 


8 parts 


Water to 


1000 parts 



Uranium ferrocyanide has the peculiar property of adsorbing basic dyes in 
direct proportion to the image density. The blue-toned image does not have 
this property, hence upon treatment with solutions containing safranine, 
chrysoidine, rhodamine B, auramine, etc., it becomes possible to modify the 
color of the red-toned image to any desired degree, without affecting the other. 
The silver of the image becomes converted into white translucent silver 
chloride, since chloride ions are present in both toning solutions. To make 
the picture more brilliant, it becomes desirable to remove this, a procedure 
which is accomplished by treatment with a hypo bath, such as: 

Hypo 904 parts 

Sodium bisulphite 22.6 parts 

Chrome alum 16.8 parts 

Potassium alum 16.8 parts 

Potassium iodide 21.6 parts 

Water to 3400 parts 

This concoction contains two mordants for basic dyes, chrome and potassium 
alum. The presence of potassium iodide would indicate that the first step 
would be the conversion of the silver chloride to the corresponding iodide, a 
substance utilized by Traube as an adsorbent of basic dyes. But it is open to 
doubt whether the silver iodide, even in combination with dyes, would re- 
main untouched by so strong a hypo bath. The process is disclosed in United 
States patent 1897369 issued to W. T. Crespinel and H. K. Fairall, and in 
English patents 339323 and 3 8 4334- 

A completely integrated process for the making of prints in color by toning 
methods, was marketed about 1936 by the Defender Photo Supply Company. 



320 HISTORY OF COLOR PHOTOGRAPHY 

The process is based upon the disclosures of Snyder and Rimbach. The sepa- 
ration negatives were printed upon a special stripping paper supplied by De- 
fender. It has been determined by actual test that the conversion of the silver 
image into the nickel-di-methyl-glyoxime compound resulted in no loss of 
density, but the conversion into the cyan resulted in a loss of 15 per cent, and 
the conversion into a yellow in a loss of 30 per cent (the " Defender Chroma- 
tone Process," seventh edition, p. 19). In making the exposures, it becomes 
necessary therefore to increase the cyan and yellow printing by 15 and 30 per 
cent respectively. Masking is recommended to improve the rendition of the 
reds and greens. For this purpose, weak positives are made from the red- 
filter negative, and developed to a gamma equal to one-third that of the 
separation negatives. These are superimposed upon the negatives. When 
printing from masked separations, the bromides should be developed approxi- 
mately thirty per cent longer. The use of a densitometer makes the gauging 
of exposure very easy. 

The green-filter negative is placed in the enlarger, and the instrument set 
for the desired size image. It is then locked in place. The negative is replaced 
by a negative of a gray scale, and a print made with a guess timing. This 
is processed to the final stage, including toning with di-methyl-glyoxime. 
The strip is then placed upon a white surface, and the step is noted where a 
pure white still persists. Let us suppose this to be the step which has a 
density of 1.50 in the negative gray scale. A point is picked out in the negative 
which would correspond to a white in the final picture. This is read. Suppose 
the reading is 1.15. There is a difference of 0.35 between the two densities. 
Since the density in the negative is lower than the one in the gray scale, less 
exposure must be given in order to preserve a pure white. The antilogarithm 
of 0.35 is 2.24. This is the factor by which the guess timing must be divided 
in order to obtain a pure white, so that if the original exposure was 150 seconds, 
the desired exposure would be 150/2.24 or 62 seconds. Let us suppose now, 
that the corresponding points in cyan and .yellow printers are 1.25 and 1.05, 
respectively. The density differential between 1.15 and 1.25 is 0.10, the anti- 
logarithm of which is 1.26. Since the cyan negative is heavier than the other, 
more exposure will be required, so that 62 must be multiplied by 1.26 to obtain 
the correct value. This gives an exposure of 78 seconds. To this must now 
be added the 15 per cent to compensate for the loss involved in toning. The 
final exposure will then be 78 X 1. 15 or 90 seconds. 

The yellow negative has a density of 0.10 less than the standard. Therefore 
the exposure time will be less by a factor of 1.26. The correct value was 
62/1.26 or 50 seconds. But on top of this must be added the thirty per cent 
required by the toning technique. The final exposure will be 50 X 1.30 or 
65 seconds. Therefore the magenta will require an exposure of 62 seconds, 
the cyan 90, and the yellow 65. 

After exposure, the bromides are developed in 55-D for i£ to 2 minutes at 
70 F if unmasked, and up to 2§ minutes if masked separations were used: 



CHEMICAL TONING 


55-D 




Metol 


2.4 parts 


Sodium sulphite 


36.0 parts 


Hydroquinone 


10 parts 


Sodium carbonate, anhydrous 


36 parts 


Potassium bromide 


4-13 parts 


Water to 


1000 parts 



321 



For use, dilute with two parts of water. Normally it would be preferable to 
use the larger quantity of bromide, since that would insure fog-free develop- 
ment. The bromides are fixed for at least five minutes in an acid hypo bath, 
which contains no alum: 

Hypo 250 parts 

Potassium metabisulphite 25 parts 

Water to 1000 parts 

After fixation, the prints are placed in water and allowed to stay until the film 
strips off the paper base. If this is not forced, smooth non-curling tissues 
result. After the tissues are stripped, they should be washed in not less than 
six changes of water. At this stage they can be toned. 

As indicated above during the discussion of the Snyder and Rimbach dis- 
closures, the magenta and cyan tissues are bleached in the same bath. This is 
prepared by adding 25 parts of the solution marked Red and Blue Toner A, 
to 25 parts of water, then add 5 parts of the standard potassium ferricyanide 
solution. This is made by dissolving 11 parts of the crystals in 100 parts of 
water. The time of treatment is fifteen minutes, after which they are washed 
again through six changes of water. In the meantime, the yellow tissue is 
bleached in its bath compounded by adding 25 parts of the solution marked 
Yellow Toner A-No. 1, to 25 parts of the solution marked Yellow Toner A-No. 2, 
then adding 5 parts of standard ferricyanide solution. The addition of toner 
No. 1 to toner No. 2 should be accompanied by vigorous stirring, to prevent 
the precipitation of cadmium salts. In this solution, the yellow tissue is 
bleached for ten minutes. It is removed, washed once, during which time 
10 parts of a 7 per cent hypo solution are added to the above, and returned for 
a further treatment of three minutes. At this time the bleach solution is 
discarded, and the tissue is washed through five changes of water. 

To convert the bleached images to color, the washed magenta tissue is 
treated with the solution marked Red Toner B, until no further change is ob- 
served. The cyan image is treated with the contents of the bottle marked 
Blue Toner B, and the yellow tissue, with the contents of the bottle marked 
Yellow Toner B. The yellow image is now complete, and after five changes 
of water, is ready for assembly. The magenta-toned image is fixed in 7 per 
cent hypo for three minutes, and is then washed in six changes of water, rinsed 
in 1 per cent acetic acid, and washed again. This completes the magenta 
image. The blue image is treated with dilute hydrochloric acid, washed in 



322 HISTORY OF COLOR PHOTOGRAPHY 

five changes of water, fixed in 7 per cent hypo for five minutes, then washed 
again in five changes of water. All three tissues are now ready for assembly. 

Since the. yellow image is the most opaque, it is placed on the bottom of the 
pack. A piece of gelatinized paper is soaked thoroughly until it has stretched 
to a maximum. It is then placed gelatin side up, upon a sheet of plate glass 
or masonite board. The yellow image is placed upon this (after a pool of 
water has been thrown on the surface), and squeegeed down thoroughly. 
The operator can mount the images either gelatin side up, in which case a 
mat surface is obtained, or collodion side up, which will give a glossy surface. 
After a few minutes to allow the surfaces to cohere somewhat, the magenta 
image is placed on the yellow. If a pool of water is placed on the surface of 
the yellow, the magenta tissue will slide very easily over it. After registry is 
achieved, the tissues are again squeegeed together, and the entire procedure 
repeated with the cyan. 

After assembly, the print is allowed to set for a while, then the edges are 
taped down firmly to the board or glass with gummed tape that is at least 
one or one and one-half inches in width. Several layers of tape should be used 
because a terrific pressure will be exerted upon the sides during drying. This 
should require eight to twelve hours, at the end of which time the print may 
be cut away from the board and mounted in the usual manner. 

A novel toning procedure which should be capable of extensive generalization 
was outlined by Dr. Bela Gaspar, of Gasparcolor fame, in one of his earliest 
contributions to color photography (U.S.P. 1956017, Re-issue 21513; Eng. P. 
379679). The main idea is that the silver image is converted into an oxidizing 
agent which is then made to act upon the ethers or ester salts of the leuco vat 
dyes. We will have more to say concerning the chemistry involved, in the 
chapter dealing with color development. 

In the specifications, Dr. Gaspar points out that it has already been proposed 
to use leuco dye bases in emulsions, and these can be oxidized to the dye in 
situ with the silver image. But the leuco bases are very unstable substances. 
The emulsions containing them could not be stored long, without air oxidation 
taking place. These difficulties could be overcome by the use of ethers, esters, 
and ester salts of the leuco bases. To oxidize these to their respective colors, 
the silver image is transformed into a compound which will act as an oxidizing 
agent. To make a blue picture, leuco indigo derivatives may be used. The 
substance di-acetyl-leuco-indigo is insoluble in water, hence may be added 
to the emulsion. It will be unaffected by alkaline developers, and neutral 
fixing baths. 

The plate containing this substance dispersed in the emulsion, is exposed, 
developed, fixed, and washed in the normal manner. This will yield a silver 
image imbedded in gelatin containing di-acetyl-leuco-indigo. The image is 
next converted into lead chromate, which may be accomplished in the following 
manner. Treatment with a ten per cent solution of potassium ferricyanide 
will oxidize the silver to silver ferrocyanide. No acid can be present, since 



CHEMICAL TONING 323 

this would saponify the acetyl groups and form the dye with no regard to the 
silver image. After a thorough wash, the image is treated with lead salts to 
form lead ferrocyanide. Another thorough wash removes excess lead ions, 
after which, treatment with a ten per cent solution of potassium chromate will 
convert lead ferrocyanide into lead chromate, a yellow pigment. If this be 
treated with dilute hydrochloric or sulphuric acid, chromic acid will be re- 
leased 

PbCr0 4 + 2HCI -> PbCl 2 + H 2 Cr0 4 

This is a strong oxidizing agent, and it will act upon the leuco derivative to 
saponify and oxidize it to the blue dye indigo. The untreated leuco base can be 
removed by treatment with acetone and hypo solutions. Instead of leuco 
indigo, the alkaline salts of the sulphuric acid ester of indoxyl can be used. 
These substances have the structures: 

Di-acetyl-leuco-indigo Sodium indoxyl sulphate 

It is also possible to use oxidized forms of the indigos, such as di-acetyl- 
indigotine. Under these conditions the silver image must be processed to 
yield one consisting of ferrous salts, after which it is made to react with the 
indigotine to yield indigo. In the first example dye formation took place by 
a procedure which involved saponification and oxidation. In the second 
example there was utilized saponification and reduction. 

In another example, the silver image is converted into one composed of 
manganese dioxide. The Gaspar specifications do not describe how that may 
be accomplished. For this step we must go back to the disclosures of the 
Neue Photographische Gesellschaft (Eng. P. 18370/03, 10898/04 and 21584/- 
06). A silver image is treated with: 

Potassium ferricyanide, 0.5% 100 parts 

Manganous sulphate, 2 % 20 parts 

Potassium bromide, 10% 15 parts 

or 

Potassium ferricyanide, 0.5% 100 parts 

Manganous sulphate, 2% 25 parts 

After the silver is completely bleached, the print is washed, then converted 
into the manganic dioxide by treatment with: 

Potassium ferricyanide, 2 % 9 parts 

Sodium hydroxide, 4% 10 parts 

This technique was evolved by the Neue Photographische Gesellschaft for the 
formation of colored images by the subsequent action of these solutions upon 



324 HISTORY OF COLOR PHOTOGRAPHY 

aromatic amines and other color formers, but it can also be used to tone 
prints by the Gaspar procedure. If the manganese dioxide image be treated 
with acids, manganic acid will be released, which could then act upon the 
sodium salt of the sulphuric ester of leuco-thio-indigo to form a red image. 
The yellow could be formed from helindon yellow. 

Vanadium salts can act as an oxygen carrier or as a catalyst for the oxida- 
tion of the leuco derivatives of the vat dyes. The formation of a vanadium 
ferrocyanide image has been discussed in some detail, above. When such an 
image, imbedded in gelatin containing the leuco derivatives, is bathed with: 



Potassium dichromate 


0.1 part 


Sodium chlorate 


0.1 part 


Sulphuric acid, 20% 


1 part 


Water to 


50 parts 



the leuco base will be saponified and simultaneously oxidized. 

The sulphate esters of the leuco vat dyes possess a strong affinity for gelatin, 
so that under normal conditions no fixing agent is required. Where this 
affinity fails, they can be pigmented with cinchonine or other alkaloids. The 
affinity for gelatin can be increased if tartaric acid or tanning agents such as 
chrome alum, are present. The sulphate esters of the vat dyes are also known 
as indigosols. 

The I.G. also disclosed a toning procedure which utilized the indigosols 
(U.S.P. 1945658; Eng. P. 365661). The silver image is treated with the 
following solution: 

Ammonium persulphate, 5% 15 parts 

Sulphuric acid, 10% 2 parts 

Copper sulphate, 10% 1-2 parts 

Potassium bromide, 10% 5-20 drops 

Indigosol pink J R, 2% 5 parts 

Water 50-70 parts 

The time of treatment is four to six minutes. The claim is made in the specifica- 
tions that the milder oxidizing agents, such as ammonium persulphate, etc., 
will not attack the indigosol even in acid solution. The silver of the image 
acts catalytically to cause this oxidation. Perhaps the true reaction is the 
release of nascent oxygen by the interaction of ammonium persulphate and 
the silver image, a fact first noticed by R. E. Liesegang {Brit. J. Phot., Vol. 44 
(1897), p. 774, 814; Vol. 45 (1898), p. 2, 646). If such is the case, then the 
solution used in the formation of gelatin reliefs by the etching process (cf. 
chapter on Transfer Processes) could be used instead of the one given above. 
The I.G. disclosed another application of dye chemistry to toning, in a 
series of patents (U.S.P. 1963197, 2127257 and 2179228; Eng. P. 387197). 
The silver image is converted into silver diazotate, and this is made to undergo 
a coupling reaction to form a dye. The chemistry of the amines and diazonium 



CHEMICAL TONING 325 

salts will be discussed in detail in the chapter dealing with diazotype photog- 
raphy, so here only a few highlights will be referred to. 

The diazonium salts are formed by the action of nitrous acid upon aromatic 
amines. 

<^ ^NHi + 2HCI + N*NO t » <j ^>H,.CI + 2Ha. O + NclCI 

Aniline Benzene diazonium 

chloride 

When treated with amines or phenols, in neutral or slightly alkaline media, a 
condensation occurs, by which a dye is formed 

< >-N a -Cj -f-H-<^3' OH > <f~3" NaBsN <^>" OH 

ft****/ 

The number of dyes that can be formed runs into the millions, hence practically 
any shade or hue can be obtained. 

Among other reactions of the diazonium salts, is one in which a metal- 
diazotate is formed, 

<( )>-N 2 CH-2NaOH > <( )>-N~N-0Na+H 2 0+NaCl 

Sodium diazotate 

The diazotates are characterized by a dynamic tautomerism, of the syn-anti 
form 

<(~~">-N«N-ONa >< >-N ;=* <(~y>-N 

NaON N-ON<i 

Syn diazotate Anti-diazotate 

These two forms can be distinguished from each other by the fact that the 
anti-diazotate does not couple, while the syn form does. The sodium diazo- 
tates are quite soluble, but the silver salts, especially of para nitro compounds, 
have but little solubility. Therefore it becomes possible to convert a silver 
image to one which is composed of silver-anti-diazotate. To accomplish this, 
the silver is first converted into silver chloride by well known methods. It is 
then treated with sodium anti-diazotates. The excess is washed out. The 
reason for the use of anti-diazotate rather than the syn form is to prevent 
any coupling with gelatin. After the image has been completely converted, 
it is treated with the proper coupler and at the right pH. The anti-diazotate 
becomes converted into the syn form and couples to form the dye. 

In another variation, aromatic amines with free — SH, — C=CH, or =NH 
groups are used. These form insoluble silver salts. Thus an image composed 
of silver bromide, is bathed for five minutes in the following solution: 

6-amino-2-mercapto-benzthiazol 1.0 part 

Sodium sulphite 1.0 part 

Sodium hydroxide 0.1 part 

Water 10b parts 



326 HISTORY OF COLOR PHOTOGRAPHY 

In this solution will be formed an insoluble image composed of silver-mer- 
capto-6-amino-benzthiazol 



"0>" 

It is washed for fifteen minutes in water containing some sulphite, and is then 
converted into the diazonium salt with acid nitrite. 

AS-S-C^ I I +NaN0a+2HCI >H 2 0+NaCl+Ag-S-C^ [ J 

This will remain anchored because it is in the form of a silver mercapto salt. 
Upon treatment with the proper coupling agent, dyes are formed. 

Images composed of azo dyes are formed by still another method by F. Lierg 
(U.S.P. 1758572). Here the silver image is converted by devious methods 
into an image consisting of complex nitrite salts. These can be made to react 
with amines to form diazonium chlorides, which are then coupled to form dyes. 
The silver image could be treated with thallic chloride. This will form thallous 
and silver chloride. Upon treatment with sodium-cobalti-nitrite, there w 7 ill 
be formed cobalt-thallium-nitrite complexes which are insoluble. Treatment 
with acids will release the nitrite. 

The dye that is formed should be as insoluble as possible. This is accom- 
plished by use of 2:3 hydroxy-naphthoyl-anilides 



m 



6 

These compounds can couple in but one position, the one that is ortho to the 
hydroxyl group. There is evidently formed the hydrozone form of the dye, 
since despite the presence of a free OH group, no solubility in alkali is obtained. 
The structure of the dye can therefore be written: 


The shade of the dye can be varied at will by the proper choice of amine. 




c^-R *° 



CHAPTER 21 
DYE TONING 



A 



CLEAR piece of film, when bathed in solutions of basic dyes, can be 
washed clear of all coloring matter. But if the film be given a previous bath 
in chrome or potassium alum, it will no longer be possible to wash the dye out. 
The chromium or aluminum ions unite with the gelatin to form stable salts or 
complexes. These can unite with the dye to form insoluble or "laked" pig- 
ments. Many other substances beside alumed gelatin, have this property of 
mordanting dyes. Since it is so easy to transform the silver of a photographic 
image into a variety of other insoluble salts, it is apparent that a further re- 
action with dye becomes possible, provided a correct choice has been made 
with regard to the conversion. This is the fundamental principle underlying 
the dye-toning processes. Slightly more general in scope, since they are not 
limited by the previous formation of a silver image, are the mordant processes. 
Here the mordant images are formed either from silver or by exposure of a 
non-mordanting light-sensitive system, which forms mordants or products 
that can be converted into mordants, by the action of light. 

Dyes can be classified into a number of groups depending upon their chemical 
behavior, reaction with fiber, chemical constitution, etc. A dye is called 
direct, if it is absorbed by the fiber without the aid of a mordant. A chrome 
dye requires the fiber to be previously treated with chromium salts before a 
true staining can occur. Among other classifications are those of acid and 
basic dyes. An acid dye is one whose chemical constitution is such that it 
unites with a base to form a salt. Such dyes usually contain an acid group 
like — COOH, — S0 3 H, — OH, etc., which contain a hydrogen that can 
ionize. With bases, these form metal salts which ionize in solution to yield 
a negatively charged dye, and a positively charged metal ion. A basic dye is 
one which unites with an acid to form a salt. These usually contain the basic 
group — NH2, or — NR 2 , where R is a hydrocarbon residue. By virtue of the 
presence of the amino group, such dyes form ammonium-like salts which 
ionize in solution to yield a positively charged dye, and a negatively charged 
acid ion. 

When an acid unites with a base, neutralization occurs. In chemistry, this 
means that the hydrogen ions of the acid unite with the hydroxyl ions of the 
base to form a substance which does not ionize readily, and which yields 
equivalent quantities of the two ions when dissociation does occur. An 
analogous reaction takes place when acid and basic dye solutions are mixed 

327 



328 HISTORY OF COLOR PHOTOGRAPHY 

in equivalent amounts. It is not necessary for the acid and basic dyes to be 
present in the forms which yield hydrogen and hydroxyl ions. Thus in a 
solution of methylene blue hydrochloride, no hydroxyl ions are present, the 
methylene blue ion being in the form of an ammonium salt. Similarly in a 
solution of acid fuchsine, the hydrogen has been replaced by a metal ion such 
as sodium. When equivalent quantities of the two dyes are mixed, the dye 
ions unite with each other (if the solution is maintained at a predetermined pH) 
to form a new unit that is electrically neutral, and generally insoluble. The 
new unit will be formed by the union of the positively charged acid dye ion 
with the negatively charged basic dye ion. 

It may be wondered why the metal or acid salts of the dyes do not act in 
the same manner. After all, the sodium salt of a dye like acid fuchsine, is also 
a union between the negatively charged dye ion and a positive sodium ion. 
But now colloidal chemistry comes into play. It is an open question whether 
a heavy dye molecule can exist in solution, except in the form of a colloid. 
This is a system in which the special properties of the particular chemicals 
involved become modified perceptibly by the surface forces which arise from 
the interaction between large surfaced particles, and the medium in which 
they are dispersed. The charge arising specifically from the dissociation of 
the dye molecule into dye and metal ion, may be translated to the dye particle 
as a whole, so that a solution of a dye can be represented as a dispersion of 
charged particles in water. The mixture of two dye solutions, one containing 
particles with a net positive charge, the other with a net negative charge, 
will cause an interaction between the heavy ions which will result in a neutrali- 
zation of the charges, and mutual precipitation of the particles. This principle 
is utilized in the determination of the strength of dye solutions. 

Many other substances dissolve in water with the formation of heavy particles 
that carry charges. Tannic acid, for instance, will act in this manner and in 
solution, if not present in large excess, will precipitate a basic dye. It may 
be stated generally that the precipitation of a dye from solution, can be 
accomplished by the addition to the solution of an oppositely charged colloid. 

When a substance is dissolved in water, the solution can be either crystalloi- 
dal or colloidal in nature. In crystalloidal solution, the dissolved particles are 
present in no larger than molecular state. If the dielectric constant is suffi- 
ciently high, and if the valence bonds in the dissolved molecule are sufficiently 
loose, dissociation will take place, with the formation of positively and nega- 
tively charged ions. The surface charges on the particles arising from such 
dissociation must be fairly high, since the particle size is very small. But 
under other conditions, it is possible that the dissolved particles exist in the 
form of conglomerations of many molecules. This gives rise to very large 
particles, upon which charges may reside due to external conditions, such as 
adsorption of ions, etc. But now the intensity of surface charge is relatively 
low, since a given charge becomes distributed over a large area. When the 
intensity of the adsorbed charges reaches a sufficiently high level, the particles 



DYE TONING 329 

maintain themselves in a state of suspension by virtue of mutual repulsion 
of like charges. Such particles are called peptized. 

It is possible to peptize many insoluble substances to cause them to go into 
colloidal solution. Silver iodide, cuprous iodide, and the corresponding 
thiocyanates are all very insoluble. Upon the addition of large excesses of 
the negative ion, adsorption upon the surface of the insoluble particles takes 
place, with the formation of either molecular or colloidal complexes that are 
highly charged. Since it is the acid ion which is adsorbed, the molecule takes 
on a net negative charge. When such particles, either in colloidal solution 
or coarse dispersion, are added to a solution of a basic dye, neutralization of 
charges takes place, and mutual precipitation results. The precipitation of 
dyes by the alums, tannic acid, and all the other mordants can be explained 
by this theory. The failure of a substance to mordant the dyes whose charges 
are not opposite to that carried by itself, is a very strong argument in favor 
of this colloidal theory of dyeing. It is only fair to state that many other 
explanations have been offered, but for the purposes of dye-toning none of the 
others appear to be so generally satisfying. An explanation somewhat along 
these lines was proposed by E. R. Bullock of the Eastman Kodak Company, 
{Trans. Far. Soc. y (1923) p. 327). 

The ability of various insoluble metal salts to adsorb dyes was tested by 
A. B. Clark (Abr. Sci. Pub., Kodak Research Laboratory, Vol. 2 (1915-1916), 
p. 130). Suspensions of many metal ferri- and ferrocyanides, and the silver 
halides were treated with dyes, and the staining properties of the dyes were 
noted. The dyes were both of the acid and basic varieties. The silver halides 
were found to be very poor in this respect. The halides of manganese, cobalt, 
nickel, and iron also were poor mordants. Practically the only dye which 
was absorbed by these halides was alizarin, a dye which is known to form 
lakes very easily with heavy metal ions. Somewhat better were the corre- 
sponding salts of bismuth, cadmium, eerie, cuprous and cupric, stannous and 
stannic, vanadium, and uranyl ions. The heavy metals like mercury and 
thorium gave the strongest action. Abnormal in its action was zinc, which, 
being a relatively light metal, should have but little attraction for the dyes, 
but actually has a strong affinity for them. Clark concluded the following: 

1. When the ferricyanide ion is present in slight excess, all ferricyanides 
mordant basic dyes better than acid dyes; 

2. Under like conditions, with excess ferrocyanide present, the same is true 
with all ferrocyanides except lead and thorium; 

3. Ferricyanides mordant better than ferrocyanides. 

It is rather peculiar that these findings have found practically no application. 
In fact the practice has been to use silver iodide as the mordant image. Under 
proper conditions it is possible to prepare this salt in a form whose absorption 
of basic dyes cannot be matched by any other type of mordant. This is the 
form in which it is utilized in the Brewster Color process. 



330 HISTORY OF COLOR PHOTOGRAPHY 

The first efforts with mordants did not utilize a silver image, but one com- 
posed of the reduction products of dichromate. The action of light upon 
dichromate and its application to photography was first observed by Mungo 
Ponton, a Scotchman (New Phil. /., (1839), p. 169). The first use of this 
product as a mordant for coloring matter came sometime in the late eighteen- 
fifties, and the credit can be divided between Testud de Beauregard and 
Masson (Photo. J., Vol. 2 (1855), p. 197), Hunt, Perry (who took out a patent 
on February 6, 1857), and V. Joseph Sella, who, in a letter published in Pho- 
tographic Notes (Vol. 2 (1857), p. 242), suggested that it may be possible to 
treat a plate that contains an image of reduced chromium salts with logwood, 
madder, sumac, etc. These are dyes, and the inference is that they would 
become mordanted upon the reduced chromium salts, and in that way in- 
tensify the image. But after that, the letter goes on to describe a process 
that was quite similar to that of Perry. 

Perhaps a clearer exposition of the principle of dye toning was contained 
in the article by E. Kopp (Phot. News., Vol. 8 (1864), p. 147). In this paper 
Mr. Kopp discussed the use of a mixed salt of ammonium and potassium 
dichromate, as the light-sensitive element. After exposure the plate was to 
be washed. Three alternative methods of intensifying the image were then 
proposed. First, if the wash is sufficient to remove all the dichromate, but 
not sufficient to cause the brown image to disappear, it becomes possible to 
treat it with salts of lead, silver, bismuth, mercury, etc., to form the corre- 
sponding opaque and highly colored chromates. Secondly, it is possible to 
treat the image with aniline, naphthylamine etc., and form intense colors by 
the action of the chromic oxide upon these amines. This is a clear anticipation 
of dye formation via the oxidation of amines, now utilized in color develop- 
ment. Third, if the wash has been very complete, it is possible to treat the 
image with logwood, alizarin, Brazil wood, purpurine, etc., in which case 
the coloring matter will become adsorbed upon the reduced chromium image. 

The following year Carey Lea (Brit. J. Phot., Vol. 12 (1865), p. 162) made 
the suggestion that if a silver image is treated with mercuric chloride, a 
mordant will be formed for murexide. He specifically mentioned that the 
procedure might be generalized to include other mordants and other dyes. 
Nothing much further appears to have been done for some time. In 1892 
A. Villain presented pictures to the Societe fran^aise de Photographie which 
appear to have been made by the -Kopp procedure, although instead of using 
pure dichromate as his light-sensitive medium, he used this mixed with 
ammonium vanadate. After exposure and washing, the prints were dyed 
with alizarin, isopurpurine, alizarin blue, etc. The application to three-color 
photographic reproduction appears first to have been made by G. Selle 
(Eng. P. 12517/99; Ger. P. 117134). Koenig, in his book, "Die Dreifarben- 
photographie" (1904), page 54, gives Selle's disclosure some rather undue 
attention. He says that this disclosure is really an original process, since it 
does not depend for its action upon the insolubilization of gelatin, but upon 






DYE TONING 331 

the reduction of dichromate to chromic oxide, in the presence of organic matter. 
But it is apparent that Kopp was keenly aware of this, and that he took ad- 
vantage of this very reaction when he intensified his images with coloring 
matter. 

Von Hiibl, on the other hand, disagrees with the Koenig estimate. Because 
the action of the dye upon image portions is only slightly more pronounced 
than it is upon non-image portions, he claims that some different explanation 
must be had. He states {Brit. J. Phot., Vol. 46 (1899), p. 538) that gelatin 
readily absorbs dyes without a mordant, hence there can be no question of 
the chromic oxide acting in that capacity. Rather, the dye washes out with 
greater difficulty from the hardened portions of the gelatin than from the 
non-hardened portions. 

This explanation is not wholly satisfactory. It is evident, since von Hiibl 
mentions the ability of gelatin to act as a mordant, that he had acid dyes in 
mind. These are the only ones for which untreated gelatin acts in this manner. 
But as indicated above, gelatin containing reduced chromium salts will retain 
basic dyes. Where this is true there may very well be a true mordant action. 
However the von Hiibl explanation is not to be discarded lightly. Some of the 
later patents by Ives and Kelley, would appear to be in the same category. 
The Ives disclosures would convert the silver image into one of silver ferri- 
cyanide by treatement with: 

Potassium ferricyanide 1.39 parts 

Chromic acid 0.35 part 

Glacial acetic acid 50 parts 

Water to 1000 parts 

In this solution the silver is oxidized either by the chromic acid and precipi- 
tated as the ferricyanide, or it is oxidized first to ferrocyanide by the ferri 
and then to ferri by the chromic acid. In either case, reduced chromium salts 
would be formed, and these would unite directly with the gelatin to form a 
tanned image. Basic dyes such as safranine, aurantia, malachite green, 
methylene blue, fuchsine, and chrysoidine, are used in concentrations of 0.25 
to 0.75 per cent, acidified with a little acetic acid (U.S.P. 1376940; Eng. P. 
193069). Since basic dyes are used, unquestionably there must be present 
some mordant action, due either to silver ferricyanide, which Clark (cf . above) 
has shown to be a good mordant, or to reduced chromium salts. The Kelley 
disclosures (U.S.P. 1411968; Eng. P. 160137; Ger. P. 378959), falls in a slightly 
different category since acid dyes may be used. The silver image, after de- 
velopment, fixation, washing, treatment with 10 per cent formaldehyde, and 
washing, is then treated with: 

Potassium dichromate 4.75 parts 

Potassium bromide 9.5 parts 

Copper sulphate 14 parts 

Hydrochloric acid 10 parts 

Water to 1000 parts 



332 HISTORY OF COLOR PHOTOGRAPHY 

After the image is bleached it is washed, then fixed, and washed again, before 
it is subjected to the action of the dyes. 

In a later patent (Eng. P. 193069; U.S.P. 1810180) Mr. Kelley proposed 
to use a yellow-dyed emulsion. One image was printed through the base, 
developed, and then treated with the bleach solution as above. The emulsion 
was next treated with chromic acid plus bromide to restore the sensitivity, 
and the other image printed on the surface. After development, the film 
was fixed, the lower image was dyed, and the upper silver image was toned 
to a complementary color. Since the dyeing operation was preceded by a 
fixing bath, there can be no question here that silver salts are acting as a 
mordant. Cuprous bromide, the other salt that may be formed, is not known 
to be a good mordant. However Kelley suggests that acid dyes be used, 
hence these would be acted on by the gelatin itself. The combination between 
acid dye and gelatin is a very loose one which can be destroyed by excess 
washing. Naturally the water will not have as strong an action in the tanned 
portions where it is absorbed to approximately one-tenth the extent to which 
it is in the unhardened portions. Hence the dye will remain there preferen- 
tially. 

The Ives procedure was again revived by J. Guardiola and Harmonicolor 
(Eng. P. 450877). Duplitized film, colored with basic dyes, and containing 
silver images, is treated with a solution containing chromic acid and potassium 
ferricyanide. This fixes the dye in the image portions of the film where silver 
was previously present. Of course the silver is present in the form of silver 
ferricyanide. Detracolor (cf. chapter on Monopacks) disclosed a monopack 
along these same lines. The emulsion contains stable leuco basic dyes. After 
development, fixation, and washing, the film is treated with a solution con- 
taining copper salts, chromic acid, and potassium ferricyanide. The leuco 
dyes are converted into the respective colors and are immediately mordanted 
upon the silver-copper ferricyanide formed. 

The Curtis Neotone process apparently works upon the Kelley scheme, al- 
though no formulas or details other than strict working directions are given. 
The separation negatives are printed upon stripping paper and developed 
in a modified Eastman D-76 formula. 

Metol 2 parts 

Sodium sulphite 100 parts 

Hydroquinone 5 parts 

Kodalk (sodium metaborate) 4 parts 

Water to 1000 parts 

After development, the prints are stopped in a solution containing one-half 
per cent glacial acetic acid, after which they can be further processed in white 
light. The prints are bleached in a special solution whose composition is not 
disclosed by the makers, the Thomas S. Curtis Laboratories, Huntington 
Park, California. After bleaching, the prints are washed for three minutes, 
then fixed in an acid fixing-bath that contains no hardener. 



DYE TONING 333 

Hypo 200 parts 

Potassium metabisulphite 10 parts 
Water to 1000 parts 

After fixation, the films are dyed in special baths which must also be obtained 
from the Curtis organization. The dyeing time is five minutes. Care must 
be taken that the tissues do not fold over, as dye transfer will take place 
wherever two dyed gelatin areas touch each other. This indicates that no 
true mordant action takes place. Before registering, the dyed tissues must 
be washed in water at 70 degrees F, to clear the highlights. At higher tem- 
peratures, the dye will wash out of the densities; at lower temperatures, too 
long a time will be required for clearance to take place. These are conditions 
typical of the Kelley type of pictures. Registry of the three images completes 
the process. 

J. M. Blaney suggested stannic ferrocyanide as a mordant (U.S.P. 133 1092) 
The image was first treated with 



Hydrochloric acid 




2 parts 


Glycerin 




75 parts 


Water to 




1000 parts 


ted into stannic ferrocyanide 


by treatme 


Oxalic acid 




4.7 parts 


Ammonium nitrate 




2 parts 


Glycerin 




50 parts 


Stannic chloride (sp. gr. 


1.50) 


13.9 parts 


Potassium ferricyanide 




3.5 parts 


Water to 




1000 parts 



The silver chloride formed in this reaction is removed by fixation with acid 
hypo. 

Iron salts have been suggested a number of times, especially in connection 
with textile fibers and cloth. The first to use this mordant appears to have 
been J. Mercer, according to W. H. Harrison (Brit. J. PhoL, Vol. 42 (1895), 
p. 55). He toned monochrome prints to various shades. The blue-green iron 
image was decomposed by treatment with lime, and the iron hydroxide thus 
formed was able to fix a large number of vegetable dyes. Mr. Mercer ex- 
hibited prints made this way in 1858, a long time before color photography 
was a practical reality. It should be recalled that the fundamental disclosures 
of Ducos du Hauron had not been made at that time. Even then the use 
of iron hydroxide was well established, for a patent disclosing its use was 
issued to J. Perry in 1856 (Eng. P. 1983/56). This is the same Perry who 
questioned so bitterly the priority of the process disclosed by V. Joseph Sella 
(cf. above). In 1864 A. Baudesson and P. Houzeau (Eng. P. 2526/64) also 
disclosed the use of iron as a mordant. More than a generation later, Stew- 
art J. Carter (Brit. J. Phot., Vol. 45 (1898), pp. 445, 449 and 806) used an 
iron printing process, which he then converted into the iron hydroxide and 



334 HISTORY OF COLOR PHOTOGRAPHY 

toned with dyes, or formed inks with gallic acid. Fabrics were sensitized with 
ferric ammonium citrate and potassium ferricyanide. Upon exposure and 
washing, this gave a blue-green image. Treatment with alkali decomposed 
the Prussian blue to yield red iron hydroxide, which then acted as a mordant 
for such dyes as resorcine green, logwood, alizarin, etc. 

Others who suggested similar processes were A. F. Hargreaves (Eng. P. 
25043/98); F. Dommer (Ger. P. 1 14923; Fr. P. 281659); an d J- Ephraim 
(Ger. P. 166832). L. F. Douglass applied the method to a silver image which 
he first toned blue-green by treatment with a standard blue-toning solution 
(U.S.P. 1450412; Fr. P. 450412), The blue-green image was then converted 
into one consisting of ferric hydroxide and toned to any desired color by use 
of basic dyes. Somewhat different in technique was the procedure outlined 
by J. I. Crabtree of the Eastman Kodak Company (U.S.P. 1389742). The 
silver image was first converted into one consisting of silver ferricyanide, by 
treatment with : 

Potassium ferricyanide 20 parts 

Potassium permanganate 20 parts 

Water to 1000 parts 

or 
Potassium ferricyanide 71 parts 

Chromic acid 29 parts 

Water to 1000 parts 

After a thorough washing, the image was treated with: 

Ferrous sulphate 50 parts 

Hydrochloric acid 50 parts 

Water to 1000 parts 

In this bath, the silver ferricyanide was converted into ferrous ferricyanide 
and silver chloride, which could be removed by fixation. By treating with 
alkali, the iron is converted into ferrous hydroxide which can be toned with 
dyes. The iron hydroxide can be removed with oxalic acid. 

The mordant action of a uranium-toned silver image was noted in the 
previous chapter. Since this substance yields an image that is so deeply 
stained, it is limited in its use to the formation of the orange red that is re- 
quired in two-color work. Silver sulphide, as a mordant, was suggested by 
Michael Robach {Brit. J. Phot., Vol. 70 (1923), p. 363; Vol. 71 (1924), p. 183). 
The silver image was first converted into a sepia by the action of two baths. 
The first was the following: 



Ammonium dichromate 


2.6 parts 


Sodium chloride 


5.2 parts 


Hydrochloric acid 


5.2 parts 


Water to 


1000 parts 



It converted the image into silver chloride. The second, a one per cent 
solution of sodium sulphide, converted the silver chloride into the sepia silver 



DYE TONING 335 

sulphide. A clearing bath after the dichromate was advised. This consisted 
of a 3 per cent solution of sodium sulphite. After a thorough wash, the sepia- 
toned print could be dyed a red with: 

Rhodamine B 5.6 parts 

Chrysoidin Y extra 0.93 part 

Auramin 9.3 parts 

Oxalic acid 3 parts 

Acetic acid 50 parts 

Water to 1000 parts 

Instead of using sodium sulphide for the formation of the sepia tone, Mr. Ro- 
bach suggests either of the following: 

Potassium sulphurata 1.5 parts 

Hypo o.3-o-75 part 

Water to 1000 parts 

or 
Sodium sulphantimoniate 0.75-1.5 parts 

Ammonia 0.3 part 

Water to 1000 parts 

This disclosure created some discussion. A. Seyewetz {Bull. Soc. franq. 
Phot., Vol. 12 (1925), p. 204) reported that his attempts to duplicate the 
Robach findings ended in failure, and he concluded that silver sulphide had 
no mordant action toward basic dyes. Mr. Bullock in " Chemical Reactions 
of the Photographic Latent Image," p. 113, states as his opinion that silver 
sulphide, prepared in an excess of soluble sulphides, is not a mordant for basic 
dyes, but that in this case the mordant action may be sought in associated 
molecules. Chromium salts must be present, and also antimony sulphide if 
sulphantimoniate were used. In a footnote to this discussion, Mr. Bullock 
pointed out that it is generally possible to convert a silver image into one that 
is composed of a negatively charged colloid, in which case it would mordant 
basic .dyes. A silver ferrocyanide image bathed in 0.1 per cent tin, antimony, 
or arsenic sulphide, kept in solution by the presence of a slight excess of am- 
monium sulphide, will form a colloidal image that will mordant basic dyes. 
In his paper before the Faraday Society (cf. above) Mr. Bullock pointed out 
that silver cyanide is much more soluble in normal cyanide than the iodide 
is in normal iodide solution. This in turn is greater than the bromide. Under 
similar conditions, the sulphide is the least soluble of the four. From these 
facts it can be concluded that silver cyanide forms the complex with the 
largest charge, with the iodide being next in intensity, and the sulphide least. 
The mordant action should follow in the same order. Therefore both Mr. Ro- 
bach and Dr. Seyewetz are correct. 

Copper salts form very potent mordants. The first use of this salt as a 
mordant for dyes appeared to have been made by J. Helouis and C. de Saint- 
Pere (Fr. P. 247065) in 1895. Fabrics were sensitized by bathing in a solution 
containing dichromates, and uranium, iron, lead, and copper salts. Under 



336 HISTORY OF COLOR PHOTOGRAPHY 

the action of light these gave the "ous" salts, which acted as mordants for 
cochineal, alizarin, the tannins, logwood, etc. R. Namias published a com- 
prehensive study on the use of copper and other metals as mordants, in 1909 
{Brit. J. Phot., Vol. 56 (1909), Color Supp., Vol. 3, pp. 68, 91). Cobalt ferro- 
cyanide he found to be completely negative, copper very unsatisfactory, and 
lead the best of all. The image was converted into lead ferrocyanide by treat- 
ment with a solution containing lead acetate, acetic acid, and potassium 
ferrocyanide. To remove the last traces of non-image lead (this metal is held 
very tenaciously by gelatin) the washing had to be thorough, and a further 
treatment with ten per cent nitric acid was found to be very beneficial. The 
silver salt formed was removed with hypo. The lead ferrocyanide was con- 
verted into the sulphate by the action of a five per cent solution of sodium 
sulphate, containing one-half per cent sulphuric acid. Dyeing was very slow, 
several hours being required from a one per cent solution. For three-color 
work, safranine served for the magenta, auramin for the yellow, and methyl 
blue for the cyan. The opaque lead sulphate could be removed with con- 
centrated hypo solutions containing 7 per cent boric acid. Since lead hy- 
droxide is more transparent than the sulphate, the ferrocyanide could be con- 
verted into the hydroxide by treatment with alkalis. This image must be 
stained in an acid solution of the dye, preferably acetic acid. 

The Namias findings seem rather strange in the light of subsequent ex- 
perience. The action of the copper ferrocyanide formed by him must have 
been exceptionally poor, if a lead mordant, requiring several hours for com- 
plete dyeing, was much more satisfactory. It is not necessary to treat a 
mordant image for a period longer than five to fifteen minutes, and then only 
if the dye solutions are extremely weak, in concentrations that are less than 
0.1 per cent. For this reason it is rather hard to see why the German govern- 
ment rejected Traube's application for a patent (Ger. P. application T20656) 
on the ground that Namias had priority. The English and French govern- 
ments did not take this limited view and granted Traube the protection he 
merited (Eng. P. 147005, 147103 and 163337; Fr. P. 491927 and 520111). 
A later modification was allowed by the German government (Ger. P. 403428) 
in which the claim was made for the presence of excess ferricyanide ions. 
Since it is not possible to compound a bath where this substance would not 
be present in a ten-thousandfold excess over the silver, the granting of this 
and the rejection of the original claim, appear to be somewhat inconsistent. 
Traube pointed out that the copper ferrocyanide (Namias to the contrary) 
possessed a very high potency as a mordant, and consequently the images 
used should be extremely thin, but full of detail. This means that ample ex- 
posure should be given to the print, and that the development should be 
very soft. A 0.1 per cent solution of the dye required but five minutes for 
complete staining. The silver should be removed with weak hypo, which will 
not affect the copper salt. Traube commercialized this process under the 
name of Uvachrome. 



DYE TONING 337 

Traube was followed by several others who made slight changes in the 
chemical composition of the bleaching solutions. J. I. Crabtree (U.S.P. 
1305962) treated the silver image with 



Copper sulphate 


12 parts 


Potassium ferricyanide 


12 parts 


Potassium citrate 


57 Parts 


Ammonium carbonate 


6 parts 


Water to 


1000 parts 



The presence of ammonium carbonate created the danger that a slight loss 
of image might result, a fact which would be felt mainly in the highlight 
region. R. Namias (cf . chapter on Chemical Toning) pointed out that both 
cupric ferri- and ferrocyanides are soluble in solutions made alkaline with 
ammonium salts. After rinsing, to remove excess bleach solution, the material 
could be dyed in 0.1 to 0.2 per cent solutions of basic dyes, with 0.1 per cent 
acetic acid present. Suitable dyes were rhodamine 6G, thioflavine, methyl 
green, or methylene blue. The silver ferrocyanide formed could be removed 
by treatment with hypo to which was added 0.25 per cent of tannin and so- 
dium acetate. 

A neutral solution of copper sulphate and potassium ferricyanide was 
proposed by F. E. Ives. Two stock solutions were prepared. These were 
mixed in equal amounts just before use (U.S.P. 1278667 and 1300619; Eng. P. 
113617 and 113618). 



Solution A: 




Potassium ferricyanide 


5-7 parts 


Potassium citrate 


27.68 parts 


Water to 


1000 parts 


Solution B: 




Copper sulphate 


7.14 parts 


Potassium citrate 


27.68 parts 


Water to 


1000 parts 



Since the citrate is present to repress the ionization of the copper sulphate 
into cupric ions, it is hard to see why it is required in the ferricyanide solution. 
After bleaching, the image is dyed with: 



Fuchsin 


0.026 part 


Auramin 


0.026 part 


Acetic acid 


1.6 parts 


Water to 


1000 parts 



It is seen here that extremely weak dye baths are used. The presence of the 
yellow dye indicates that a two-color process is intended. In a subsequent 
disclosure (U.S.P. 1499930 and 1655182; Eng. P. 119254) Ives proposed to 
print two images in the same emulsion layer. A yellow dyed film is used. 
The red-filter image is printed through the base, giving an image that lies 



338 HISTORY OF COLOR PHOTOGRAPHY 

adjacent to the celluloid, and which penetrates but a short distance into the 
film. This is toned blue-green by means of iron. The green-filter separation 
is printed in registry, in the normal manner. This gives a surface image that 
does not penetrate deeply, so does not reach the lower one. The upper image 
is converted into a red by use of the technique just disclosed. The same 
idea is utilized by L. F. Douglass (U.S.P. 1632278); W. V. D. Kelley (U.S.P. 
1712439 and 1810180); Max B. Dupont (Eng. P. 360109); and W. H. Peck 
(U.S.P. 1840524). 

In all of the above disclosures, it is cupric ferrocyanide that is used as a 
mordant. As E. J. Wall points out in his "History of Three-Color Photog- 
raphy" (p. 371) it is perfectly possible to accomplish the same result by first 
treating the silver image with ferricyanide, then with copper salts. What 
is not pointed out, is that a tremendous loss of image density results. But 
this is not an unmixed blessing, since it allows the operator to make a more 
normal print for toning. After dyeing, the dye image may be fixed with tannic 
acid or some stronger mordant than copper ferrocyanide, and this last re- 
moved by treatment with alkali. This use of two-bath toning was suggested 
by F. J. Ventujol (Fr. P. 558699). A two-bath solution proposed by the 
Naturfarben Film g.m.b.H. (Ger. P. 393790) differs from the above in that 
cuprous ferricyanide is the mordant. This is formed by the action of cupric 
chloride upon the silver image, followed by a ferricyanide bath. 

Equal in potency to the ferrocyanide, but free from the color of this image, 
is cuprous thiocyanate, first proposed by J. H. Christensen (U.S.P. 1447759; 
Eng. P. 132846 and 135477; Ger. P. 319459* 3 J 9477 and 334277)- The silver 
image is treated with the following solution: 



Potassium citrate 


55.5 parts 


Copper sulphate 


41.75 parts 


Potassium thiocyanate 


20 parts 


Acetic acid 


25 parts 


Water to 


1000 parts 



This formula was modified to a slight extent by Seyewetz, in a paper dealing 
with mordant processes in general. Dr. Seyewetz is quite partial to the cuprous 
thiocyanate mordant. His modified formula {Brit. J. Phot., Vol. 71 (1924), 
p. 609) is the following: 



Copper sulphate 


40 parts 


Potassium thiocyanate 


20 parts 


Potassium citrate 


60 parts 


Acetic acid 


30 parts 


Water to 


1000 parts 



The bleached print is a dirty gray in color. This does not modify the shade 
of the image, although it leads to a slight degradation. An improved formula, 
in that keeping qualities are much better, was given by Lyman Chalkley. 



DYE TONING 339 

This is characterized by the presence of high concentrations of potassium 
citrate 



Copper sulphate 


40 parts 


Potassium citrate 


250 parts 


Acetic acid 


30 parts 


Ammonium thiocyanate 


25 parts 


Water to 


1000 parts 



The ammonium thiocyanate could be replaced by the potassium salt, of which 
29 parts should be used. 

This formula can be further modified to advantage to yield practically an 
invisible image (Joseph Friedman, Am. Phot., Vol. 32 (1938), p. 212). 
This is accomplished if the thiocyanate concentration is increased. It is 
possible to have a sufficiently high concentration to completely fix a silver- 
bromide emulsion. But true fixation occurs in two steps, the first of which 
is the formation of a complex silver salt whose index of refraction is the same 
as that of gelatin. At this stage the film becomes completely transparent. 
If the size of the complex salt is such that it is very mobile, it will quickly 
diffuse out of the gelatin, but if its mobility is greatly restricted, it will remain 
fixed. It is to be recalled that Ansco and the I.G. in their successful solution 
of the monopack problem utilized this principle in the preparation of their 
film. If the concentration of the thiocyanate be carefully determined, and 
if the time of treatment be controlled, it is possible to convert a silver into a 
silver-copper-thiocyanate image that is practically invisible, and one whose 
dye absorption properties are at a maximum. 

Probably the best mordant of all, and the simplest to operate, is the iodide. 
This can be either the silver or the copper salt, it being immaterial which is 
used. The first public disclosure of silver iodide as a mordant came from 
Dr. A. Traube <Ger. P. 187289 and 188164; Eng. P. 10258/07; U.S.P. 
1093503). The silver image is treated with oxidizing agents in the presence 
of the desired ion, in which case it becomes possible to precipitate either 
silver iodide, silver-copper-iodide, the corresponding bromides, chlorides, 
ferrocyanides, or the corresponding silver-mercury compounds, etc. Upon 
treatment with basic dyes, these became fixed to the bleached image. In or- 
der to obtain transparency the insoluble silver complexes must be removed, 
a procedure that could take place only after the dye had been treated with 
tannic acid and sodium acetate. The final image was not a silver-iodide-dye 
complex. Traube suggested that stripping celluloid film could be used. This 
is very interesting since Chromatone and other stripping papers take on a 
general dye stain. This is probably because the product is not celluloid, but 
collodion, and this is a mordant for basic dyes, a fact utilized by Dufaycolor 
(cf. chapter on Screen Plates). Evidently the addition of considerable 
quantities of the plasticizer camphor causes a complex formation with the 
cellulose nitrate micelle to take place, in which the negative charge becomes 



340 HISTORY OF COLOR PHOTOGRAPHY 

considerably reduced or converted into a positive charge. In this condition, 
a basic dye will not be absorbed by the base. 

Traube's claim to priority was soon contested by Tauleigne and Mazo, 
whose own disclosures came in 1909 (Fr. P. 420584; Eng. P. 27818/09; U.S.P. 
1059917). But they claimed that their first experiments were made in 1897. 
However they did make an improvement. The silver image is first converted 
into the ferrocyanide by treatment with ferricyanide. Or it can be converted 
into a mixture of cuprous and silver chloride by the action of cupric chloride. 
The bleached image is next acted on by a weak potassium iodide solution, 
2 to 3 per cent, which converts it into a mixture of cuprous and silver iodide. 
This absorbs dye better than pure silver iodide, but it is equally opaque. A 
real contribution came in the next step, which consisted in treating the dyed 
image with a solution containing 20 per cent of potassium iodide. As was 
indicated above, silver iodide is quite soluble in a strong solution of potassium 
iodide. This rendered the iodides transparent, but, as they were careful to 
note, did not remove them, thus maintaining the dye in its proper position. 
This is a very clear anticipation of the formation of the silver-iodide hydrosol 
which forms the basis of a patent issued later to Hoyt Miller (U.S.P. 12 14940; 
Eng. P. 100098; Fr, P. 483764; Ger. P. 405962). Tauleigne and Mazo did 
not realize, of course, that the transparent substance into which the mordant 
was converted was itself a far better absorbent for basic dyes, than was the 
cuprous-silver iodide. 

The Miller disclosure represents a distinct advance in mordant procedure, 
and is worthy of special attention. Unlike most patentees, Mr. Miller gives 
full acknowledgment to the earlier experimenters, Traube, Tauleigne and 
Mazo, and Thornton (Eng. P. 25084/12), and he pointed out that the first 
and last named workers formed opaque mordants that had subsequently to 
be fixed out, and that the second group dyed an opaque mordant which they 
later converted into a transparent material. Mr. Miller then discusses the 
work of Bancroft, (/. Phys. Chem., Vol. 14 (1910); and Lottermoser (/. PracL 
Chetn., Vol. 68, p. 341; Vol. 72, p. 39; Vol. 73, p. 324), where the principles 
involved in the formation of silver-iodide "hydrosols" are discussed. These 
gentlemen point out that the insoluble salt, silver iodide, absorbs potassium 
iodide, the degree of absorption being dependent upon the concentration of 
iodide ions. The silver iodide particle with the adsorbed iodide ions begins 
to go into colloidal solution, the extent depending upon the amount of iodide 
present. Therefore, it becomes possible to form colloidal solutions, or hydro- 
sols, of silver iodide with almost any desired degree of dispersion. The very 
coarse dispersions are formed in the presence of slight excesses of potassium 
iodide. These are turbid solutions. When the concentration of potassium 
iodide is sufficiently high, the dispersion becomes extremely fine, and com- 
pletely transparent: This is the form desired by Mr. Miller. 

Acting upon the ideas disclosed in these papers, and having the Tauleigne 
and Mazo example before him, Mr. Miller decided that if the silver of the 



DYE TONING 341 

image is converted into silver iodide in the presence of high concentrations 
of potassium iodide, a completely transparent silver iodide-potassium iodide 
mordant is formed. The solution he proposed was the following: 

Potassium iodide 50-70 parts 

Iodine 0.7-3.0 parts 

Acetic acid 15 parts 

Water to 1000 parts 

When bleached in this solution, a completely invisible image composed of a 
complex potassium silver iodide, is formed, an image which absorbs dye to 
a degree many times that of the opaque mordant. Since the gelatin becomes 
softened by the action of so high a concentration of potassium iodide, it 
becomes desirable to first harden the gelatin with formaldehyde. Alum, 
tannic acid, or other agents which harden gelatin and which also precipitate 
basic dyes, cannot be used. Fomaldehyde has no such action upon basic 
dyes, but the formaldehyde gelatin must first be dried before any tanning 
takes place. Not only does this procedure preserve the gelatin, but it also 
gives much improved whites, since tanned gelatin will not absorb dyes as 
readily as the soft variety. But the absorption by the image will not suffer 
materially since here there is a mutual attraction between dye and mordant. 

The silver iodide complex is completely soluble in concentrated potassium 
iodide solutions. But if it is formed so that a layer of gelatin lies between it 
and the solution, then dissolution must be followed by diffusion before it is 
removed from the locality of formation. The mobility of the complex cannot 
be very high, since the particle is colloidal in nature, and is quite heavy. To 
insure the presence of a relatively thick gelatin layer over the image, one of 
several possible techniques can be adopted. The simplest would be to print 
the image through the base of the emulsion, a procedure that is possible only 
if the base is composed of a transparent substance. Another and more useful 
possibility lies in the use of silver solvents in the developer, compounded to 
have a fairly high initial appearance time. This means that hypo should 
be added. Or potassium iodide can replace the hypo, which, as was indicated 
in the chapter on Processing Screen Plates, also drives an image deep into 
the gelatin layer. 

These are not the only considerations involved. The extreme potency of the 
complex mordant to absorb basic dyes, makes the use of normal images highly 
undesirable. The writer has at one time made a determination of the ability 
of this type of mordant to absorb dyes. To this end a gray scale was copied 
upon a lantern plate. After development, fixation, washing, and formaldehyde 
tanning, one-half of the scale was covered with rubber cement. This formed 
a water-insoluble and impermeable coating which prevented any action of the 
bleach and dye bath upon the silver beneath it. After completely processing 
the plate to a final dyed image, the rubber cement was removed. This left 
every step in the gray scale in silver and in color. The relationship between 



342 HISTORY OF COLOR PHOTOGRAPHY 

the silver and dye was determined densitometrically, by reading both through 
a filter complementary to the dye. With astraphloxine FF and rhoduline 
sky blue as the dyes, the following results were obtained, all other conditions 
being identical. 

ASTRAPHLOXINE FF 

5-minute development j-minute development 



Silver 


Dye 


Intensification 


Silver 


Dye 


Intensification 


Density 


Density 


Density 


Density 


Density 


Density 


.41 


2.30 


5.9 


.48 


2.32 


4.8 


.28 


2.07 


74 


.40 


2.21 


5-5 


.18 


1.90 


10.6 


.32 


2.05 


6.4 


.15 


1.67 


11.0 


.27 


1.86 


7.0 


.12 


1.20 


10.0 


.17 


1.56 


9.1 


.07 


0.70 


10.0 


.12 


1.22 


10.0 


.04 


0.25 


6.3 


.04 


0.68 


17.0 



RHODULINE SKY BLUE 

j-minute development 



ilver 


Dye 


Intensification 


znsity 


Density 


Density 


.50 


2.36 


4-7 


.40 


2.28 


5-7 


•34 


1.98 


5.8 


.26 


1.60 


6.1 


.20 


1.18 


5-9 


.14 


0.70 


5.o 


.03 


0.20 


6.7 



It must be kept in mind that the accuracy of the readings is very low because 
of the difficulties of heterochromatic densitometry, and the reading of very 
low silver densities. But even so some very important points stand out. 
The ability of the mordant to fix dyes is dependent upon the method of forma- 
tion. Thus, long development and short exposure give apparently better 
and more uniform intensification than long exposure and short development. 
This is apparent from a consideration of the results with astraphloxine. With 
five minutes development the intensification was almost uniformly 10.0, while 
with three minutes it varied from a low of 4.8 to a high of 10 (disregarding 
the value of 17 for a silver density of 0.04, since an error of reading at this 
point of 0.02 will be equivalent to an error of 50 per cent). 

Let us compare the results for rhoduline sky blue with those for astra- 
phloxine. The constancy of the intensification factor indicates that the dyeing 
was to approximately the same extent as in the five-minute astraphloxine 
experiment. But a generally lower value, 5.7 as compared to 10.0, gives an 



DYE TONING 343 

indication that every dye is absorbed to a different extent. This was verified 
for a number of other dyes. Other factors which affect the intensifications 
are the concentrations of the dye baths, and the concentration of potassium 
iodide in the bleach. In order to make consistent prints by this method, it 
becomes necessary to standardize the various operations very carefully. The 
writer is very partial to dye toning as a method for making prints, and believes 
that the quality of reproduction is far superior to any of the other methods. 

Since the intensification is so high (the presence of as little silver as is re- 
quired to give a density of 0.04 already yielding middle tones), it is essential 
that the development be extremely low, and the fog level of the emulsion 
equally low. For paper prints, the highest density permissible is approxi- 
mately 1.60. The silver density for a magenta image composed of astra- 
phloxine, which will yield a dye image with a density of 1.60, is 0.16. This 
therefore is the maximum silver density allowable in the black-and-white 
stage of the process. Only the very contrasty emulsions, process, super- 
process, etc., will operate at a low enough toe level to give the final image 
sufficient brightness. The contrasty lantern-plate emulsion is also suitable. 
The problem arises then, how to concoct a processing technique which will 
give straight-line reproduction in a region limited between the values 0.00 
and 0.16, which will be a true representation of densities ranging from 0.35 
to 2.0 in the original. The most successful procedure to adopt would be to 
use specially compounded developers. Those who are not susceptible to 
paraphenylenediamine poisoning, will be able to use this excellent agent. 
When compounded with glycin, as in the Sease No. 3 formula, it has fair keep- 
ing qualities 

Paraphenylenediamine 10 parts 

Sodium sulphite 90 parts 

Glycin 6 parts 

Water to 1000 parts 

If one cannot work with this chemical, he can compound a glycin formula 
by leaving out the paraphenylenediamine, and adding just sufficient alkali 
(carbonate, Kodalk, borax) to give a density of 0.16 in the deepest shadows 
in five to seven minutes development at 70 F. Another, and perhaps more 
feasible solution, would be to use metol. The basic formula is 

Metol 5 parts 

Sulphite 100 parts 

Borax 10-15 parts 

Boric acid 5-30 parts 

Potassium thiocyanate \ part 

Potassium bromide \ part 

Hypo \ part 

Water to 1000 parts 

The concentration of borax and boric acid will vary with the emulsion, but 
once established should not be changed as long as that material is being used. 



344 HISTORY OF COLOR PHOTOGRAPHY 

By properly varying the boric acid content, it is possible to make solutions 
that will require as much as one hour to yield a top density of 0.16 in the 
deepest shadows, even with as contrasty a material as the Eastman Super 
Contrast Process Plate. In an effort to compound a fine-grain developer for 
negative materials, Joseph Friedman (Amer. Phot., Vol. 33 (1939), p. 738) 
proposed the following solution: 

Metol 2 parts 

Sulphite 30 parts 

Potassium metabisulphite 30 parts 

Chrome alum 20 parts 

Hydroquinone 5 parts 

Formalin 25 parts 

Boric acid 20 parts 

Water to 2000 parts 

This gave excellent results as far as detail in very low densities was concerned, 
although the deposit was not sufficient to allow the images to be printed. It 
would appear, therefore, to be suitable for the purposes of dye toning. The 
metabisulphite, reacting with formalin, releases one equivalent of alkali, 
which is buffered by the boric acid. The chrome alum is present to prevent 
too great a swelling of the emulsion layer, and it is converted into a complex 
salt by the sulphite. Hence it will remain in solution. The hydroquinone 
probably has no action at the low alkalinity present in this solution, but it 
acts as a preservative. 

Regardless of what procedure is used to obtain the very low densities re- 
quired, it is a safe practice to use enough solution to process a single plate, 
or set of plates, and to discard it. Temperature controls are very necessary, 
at least until the operator can learn to judge the results in the dim red of the 
safelight. Since densities that are not normally even suitable highlights 
correspond to deep shadows, the safety of the light is very important. This 
should be tested by covering half the plate with a sheet of black paper, and 
exposing to the safelight for fifteen minutes, at a distance equivalent to that 
used during development. The processing of the plate should be in total 
darkness for the maximum time to be used in actual practice, say fifteen 
minutes. A short stop in one-half per cent acetic acid, followed by a rinse, 
precedes the fixation, which should be done in two stages. In the first stage 
the plate remains until the image clears. After a five-minute wash, the 
plate is again treated with fresh hypo for a further period of five minutes. 
In both cases only plain hypo, or hypo to which i| per cent of sodium bisul- 
phite or potassium metabisulphite has been added, should be used. After 
another ten-minute wash, the plate is treated with a one per cent formaldehyde 
solution and immediately dried. This can be greatly facilitated if it be first 
treated with 80 per cent alcohol containing i| per cent formaldehyde. In 
this case the drying will be rapid. No heat should be applied, since this 
would have a tendency to drive off the formaldehyde. 

After the plate is dried, it can be treated with the bleach. As stated above 



DYE TONING 345 

this is an iodine-iodide mixture containing from five to ten per cent of po- 
tassium iodide per liter of solution. Because of the solubility of silver iodide 
in potassium iodide, this solution should not be used too long. However, this 
tendency can be reduced to a minimum if the iodine solution is loaded with 
agents that will prevent the penetration of the solution into the gelatin layer. 
The diffusion of the salts out of the emulsion layer will also be greatly hindered. 
Hence, even if the silver mordant does dissolve, because of the very limited 
mobility of the substance, it will not diffuse out of the film. A suitable bleach 
along these lines, can be compounded as follows: 



Stock Iodide Solution: 




Water 


100 parts 


Potassium iodide 


100 parts 


Iodine 


25 parts 



The best method of preparing this is to make an intimate mixture of the 
iodine and iodide, then add the water in small amounts. The volume should 
then be brought up to 250 cc, making a 10 per cent solution of iodine. The 
bleach solution is prepared from this as follows: 



Water 


750 parts 


Acetic acid 


25 parts 


Sodium sulphate (anhydrous) 


100 parts 


Stock iodine solution 


25 parts 


Water to 


1000 parts 



The plate containing the silver image (fog test) is immersed in this solution 
for a time sufficient for the image to disappear. If it is correctly printed, this 
should not require longer than two or three minutes. At the end of this time, 
the plate is given a one-minute bath in a one per cent solution of sodium 
bisulphite or metabisulphite, then a two-minute wash in running water, after 
which it may be dyed. The dye baths are compounded to contain approxi- 
mately one-half to one per cent of dye in three to five per cent of acetic acid. 
If great purity is desired, the dye should be dissolved in 100 cc of alcohol, 
filtered, then made up to volume with water and acetic acid (30 to 50 cc per 
liter). The dyeing time is five minutes. The excess dye is removed by wash- 
ing in running water, and from fifteen minutes to one hour should be allowed. 
Sometimes it is possible to wash out the last traces of non-image dye with weak 
acetic acid (one per cent) or alcohol (25 per cent). 

Any of the basic dyes are suitable for this process, but the writer has had 
good success with astraphloxine FF for the magenta, rhoduline sky blue for the 
cyan, and auramin for the yellow. This last may be somewhat too far on the 
orange side, in which case thioflavine T, or some of the phloxines could be 
used. All of the dyes with the exception of auramin, can be prepared as stock 
solutions, to be used until exhausted. But this last named dye is unstable in 
solution, so must be prepared fresh each time. It has a limited solubility, so 
that its preparation is a little tricky. To prepare it in stable form, ten parts 
of the dye are dissolved in 100 parts of alcohol, the requisite acetic acid added, 



346 HISTORY OF COLOR PHOTOGRAPHY 

and the solution made up to volume with water. Instead of astraphloxine FF 
for the magenta, the rhodamines such as B or 6G, could be used. The shade 
of rhodamine B is probably the closest to theoretical requirements, but it 
does not wash out as readily as the other. Rhoduline sky blue could be re- 
placed by setopolin, thionine blue, or methylene blue, although the last named 
is the least worthy of the group. 

The best dye to use to detect whether the safelights are really safe, is astra- 
phloxine. If, after complete processing of the plate, a line of demarkation is 
distinct between the exposed and unexposed sections of the plate, then the 
safelight should be replaced. Since the dye intensifies the image ten times, 
this test is at least ten times as severe as the one ordinarily used. 

The Miller patent was assigned to the Brewster Color Film Corporation, 
and it formed the basic principle underlying the process. The entire scheme is 
disclosed in a series of patents all issued to P. D. Brewster, the earlier ones 
(U.S.P. 1308538, 1563959, 1580114 and 1580115), dealing with a two-color 
system, and the later ones (U.S.P. 1992169 and 2070222; Eng. P. 449678, 
449749 and 449750) with three-color. Since long reels of film must be processed 
at one time, the keeping qualities of the solutions are of great importance. 
The tendency of the Miller bleach to dissolve out some of the silver-iodide 
made that item very short lived. Not only did this give rise to a loss of high- 
light detail, but it muddied the whites since some of the silver iodide in the 
solution transferred to and was retained by the film. This absorbed dye as 
well as the image-bearing portions of the film. The Friedman modification, 
containing the anti-swelling or loading agents (sodium sulphate) greatly ex- 
tended the life of the solution. 

Mr. Brewster devised another scheme which accomplished the same purpose, 
and with much less danger. He achieved this result by forming the bleach 
within the image layers. The film after being tanned and dried, is treated 
with a priming solution which contains potassium iodide and sodium iodate. 
From this bath, the film is led into a gas chamber and exposed to acetic acid 
fumes. The iodate, iodide, and acid react to form iodine, in accordance with 
the following scheme, 

" NalO + 5KI + 6HOAc -> 3 I 2 + 5 KOAc + NaOAc + 3H2O 

Thus it is seen that 2.0 parts of sodium iodate will react with 8.5 parts of po- 
tassium iodide, to yield 8.0 parts of free iodine. In order to form the proper 
image, there must be present 10 per cent of potassium iodide, and 0.4 per cent 
of iodine. To develop this quantity of free iodine, there will be required 1.0 
part of sodium iodate and 4.25 parts of potassium iodide per 1000 parts of 
solution. To this must be added 100 parts of potassium iodide. The final 
concentration is as follows : 

Sodium iodate 1.0 part 

Potassium iodide 104.25 parts 

Water to 1000 parts 




DYE TONING 347 

A neutral solution of potassium iodide, when exposed to light, decomposes to 
form free iodine, which immediately reacts with water to form acid. There- 
fore, after some time, the priming bath begins to turn slightly reddish. When 
the film is treated with this solution, the free iodine, however slight, will etch 
out some of the highlight detail. To prevent this from happening, the priming 
bath is maintained at a pH just on the alkaline side, at about 7.5. If it is 
made too alkaline, the acidity of the gas will be insufficient to set off the re- 
action. A few drops of borax or sodium acetate solution, enough to give a 



cover 



contpa rfih en£ 
CO/Hf>arZhten,t~ 

FIG. 108 

definite indication of alkalinity, will suffice; or ammonia can be used. Now 
very little of the total concentration of the acid fumes will be used up to over- 
come the alkalinity of the priming bath. The use of a gas box is inconvenient 
for the technician without elaborate equipment, although a desiccator can be 
converted very easily into such. This is a piece of glass apparatus shaped 
as shown in Fig. 108. It is used by chemists to dry substances which are heat 
unstable. As they use it, concentrated sulphuric acid is placed in the bottom 
of the desiccator, in a thin layer. Calcium chloride or some other water- 
absorbing medium can replace the sulphuric acid. A perforated plate is 
placed across the indentations, and this forms a shelf upon which the material 
to be dried can be placed. To convert it into a gas chamber, a volatile acid 
replaces the drying substance. Care should be taken to allow the gas free 
access to the upper chamber. 

If this is unhandy, it is possible to put the acid, in this case acetic or some 
other equally strong organic acid, into a solvent that does not wet gelatin 
and which is not miscible with water. A substance like benzene, toluene, or 
some other liquid hydrocarbon having a limited solubility in water, is ideal. 
In that case the hydrocarbon is saturated with a concentrated solution of the 
strong inorganic acid, or some acetic or other organic acid can be dissolved 
in it. After leaving the priming bath, the film is passed through this solution. 
The tendency will be for the gelatin to absorb the acid from the hydrocarbon 
solution, rather than for the hydrocarbon to remove the primer from the gela- 
tin. Therefore the iodine will be formed in the gelatin layer, and no diffusion 
or loss of image will be possible. The further treatment of the film is the same 
as outlined above. A bisulphite bath removes the excess iodine, and a wash 



348 HISTORY OF COLOR PHOTOGRAPHY 

removes excess bisulphite. This is followed by dyeing and removal of excess 
dye, after which the film is dried. 

The Brewster process utilized a duplitized film with a yellow dyed emulsion 
on each side. This yielded two images, one of which was dyed a cyan, the 
other a magenta. In the old days when two-color was acceptable in the motion- 
picture industry, this served well. But since a revived Technicolor accustomed 
the public to the beauty of a three-color system, two-color no longer was 
satisfactory. In order to put a third color upon the duplitized film, Mr. Brew- 
ster recoated it with plain gelatin, then transferred an image of silver nitrate. 
This was immediately treated with 10 per cent potassium iodide to form the 
silver iodide hydrosol, which was then dyed with auramin. It was essential to 
put a layer of gelatin over the dyed images before transferring the silver ni- 
trate, for otherwise this substance would react with the iodide ions adsorbed 
upon the silver iodide, to form a Traube type of image. Not only would this 
give great opacity, but it would release the dye which is held by the mordant. 

This reaction could be used when it is desired to effect extreme intensifica- 
tion of a silver image. By treatment with a Miller-type bleach, the silver 
could be converted into the hydrosol, so that at least three, and possibly more 
iodide ions will be held by each atom of silver. Upon treatment with an 
ammoniacal solution of silver nitrate, the excess iodide ions are converted 
into the corresponding silver salt. A thorough wash, preferably with water 
made slightly alkaline with ammonia or with sodium sulphite, will remove the 
excess silver ions, especially those which would otherwise form silver-gelatin 
complexes (cf. chapter on Processing Screen Plates). In this manner three or 
more silver atoms will be present where only one was present previously. 
The reduction to metallic state can be accomplished by treatment with sodium 
hydrosulphite. 

The decomposition of the silver-iodide and potassium-iodide dye complex 
does not take place if copper replaces silver in the transfer. The latest Brewster 
technique makes use of this. The film, after drying, is treated with a mixture 
of copper, tannic acid, and chromium salts to completely fix the dyes in the 
film. A thorough wash removes the excess. Now there will be no interference 
between the yellow dye and the dyed image already present. Cuprous chloride 
is the substance transferred. To retain this in solution, it must be heavily 
charged with chlorides, either as hydrochloric acid, or sodium or ammonium 
chloride, or a combination of the three. It is also advisable to add some 
sodium sulphite or metallic copper, to prevent the oxidation of the copper to 
cupric salts. After transfer, the cuprous chloride is converted into the iodide 
or thiocyanate by treatment with concentrated solutions of these ions. 

Before Brewster could successfully commercialize his product, several other 
pitfalls had to be overcome. One of these is common to all color processes, 
and that is the treatment of the sound track. It is highly desirable that this 
be a black-and-white image. Brewster accomplished this by applying to the 
sound track area only, a solution which destroyed the hydrosol, reforming 



DYE TONING 349 

silver iodide as a Traube-type of image. This was blackened by treatment 
with hydrosulphite. A suitable oxidizing agent was potassium permanganate, 
sodium chloride, and an acid. Another problem was due to a defect in the 
duplitized emulsion, a defect that has now been corrected. This could not be 
detected except by the highly intensifying procedure used by Brewster. In 
order to eliminate the defect, the film was subjected to the action of bromine 
vapor, which completely removed any latent image existing in the film. Since 
a gas was used, no penetration took place into the depth of the emulsion, so 
that the destructive action was limited to the very surface. The oxides of 
nitrogen, chlorine, ozone, and other gases capable of exerting an oxidizing 
action upon the latent image, could also be used. 

The success achieved by Traube, Tauleigne and Mazo, and Brewster, soon 
centered considerable attention upon iodide as a mordant. H. D'Arcy Power 
(Brit. J. PkoL, Vol. 59 (191 2), p. 41), proposed to use a bath in which the silver 
was oxidized by ferricyanide in the presence of potassium iodide. 

Potassium ferricyanide 40 parts 

Potassium iodide 20 parts 

Water to 1000 parts 

An opaque Traube-type of image was formed, and the further treatment was 
identical with that of Traube. At approximately the same time C. Wolf- 
Czapek proposed a much weaker bath, but containing both potassium salts. 
Dye also was present so that bleaching and dyeing were accomplished simul- 
taneously. 

Potassium ferricyanide 2 parts 

Potassium iodide 3.5 parts 

Dye solution, 2% solution 15 parts 

Sulphuric acid, 10% 1 part 

Water to 1000 parts 

The dyes that were suitable for this process were auramine, chrysoidine, fuch- 
sin, methylene blue, methyl violet, rhodamine, safranin, thionin blue, and 
Victoria blue. It is interesting to note that Wolf-Czapek also claims the use 
of acid dyes, and he lists the following: tartrazine, Victoria yellow, naphthalin 
blue, acid violet, acid fuchsin, erythrosin, and rose Bengal. These procedures 
yield the opaque type of image, so that the silver iodide must be removed 
from the final result by fixation in cyanide or hypo. The dye must, however, 
be first precipitated by tannic acid before the mordant is removed. 

The Tauleigne and Mazo technique is utilized by J. H. Christensen (Eng. P. 
132846 and 135477; U.S.P. 1447759). After treating the image with cuprous 
bromide, it can be simultaneously converted into a ,silver-iodide hydrosol and 
dyed, by treatment with: 

Potassium citrate 71.5 parts 

Potassium iodide 143 parts 
Dye, to desired concentration 

Water to 1000 parts 



350 HISTORY OF COLOR PHOTOGRAPHY 

It may be a little difficult to retain the dye in solution in the presence of so 
much salt, but this can be accomplished by the addition of acetic acid. 

In the same disclosure Christensen describes a new mordant, one that is 
capable of great generalization. The ability of silver to form complex salts 
with thiourea has been well known for some time, and this and analogous 
substances have been proposed as fixing agents. Christensen disclosed that 
they may be used as mordants for basic dyes. The complexes have the same 
high transparency that is characteristic of the Miller-type of hydrosol. The 
silver of the image is first converted into silver halide, and it is then treated 
with: 

Potassium metabisulphite 7.8 parts 

Thiourea 6 parts 

Potassium iodide 3.1 parts 

Potassium sulphocyanide 15.5 parts 

Water to 1000 parts 

The plate should remain in this solution until it appears to be cleared, after 
which it is rinsed in water and dyed. This has since been repatented by True 
Colour Film (Eng. P. 466710), where the mordant is generalized to include 
all thioamino compounds, of which thiourea and allyl thiourea are typical 
examples. 

The Brewster-type of procedure was patented again by W. T. Crespinel 
and Cinecolor (U.S.P. 2016666; Eng. P. 459234 and 462353); and by A. M. 
Gundelfinger and Cinecolor (U.S.P. 2141354; Eng. P. 466290). F. Lierg 
(Eng. P. 335930) disclosed some general methods for the formation of trans- 
parent iodide and other silver-salt-complexes that would act as mordants. 
H. Shorrocks (U.S.P. 1303506; Eng. P. 111054), shortly after Miller's dis- 
closures suggested the use of a silver-iodide mordant for making a two-color 
picture. In a film containing both red and green positive images, the green 
silver deposits are protected from the action of a bleach. This converts the 
red images into silver iodide. The protection is then removed, and the entire 
film subjected to the action of a green toning bath which contains a base red 
dye. Since the green toner is chosen so as not to act as a mordant for the basic 
dye, the two act differentially. W. V. D. Kelley also used silver iodide for one 
of the images. The red negative was printed upon a positive film, then proc- 
essed to form an image in silver iodide. The film was sensitized with dichro- 
mate and the green image printed. This was dyed with a pinatype dye which 
would take only in the soft gelatin, after which the silver-iodide image was 
dyed with the appropriate color. M. S. Procoudin-Gorsky (Eng. P. 293038) 
iodized each component, then dyed in a bath containing hypo. He claimed 
that he obtained a tanned image in dyed gelatin that was free from silver. 
It is a little hard to follow this reasoning, as iodization, unless accomplished 
by a treatment with a solution containing a soluble iodide and acid dichromate, 
would not tan the gelatin. W. R. Reid and H.'V. A. Briscoe (Eng. P. 370999) 
also made individual color components which they proceeded to convert into 



DYE TONING 351 

mordant images. These were dyed, then registered upon a single base. 
P. E. F. Lessertisseux (Eng. P. 420356) used duplitized film to form a three- 
color picture. One image was printed upon each side of the film base, then 
converted into mordants and dyed. The entire film was resensitized with 
dichromate and exposed under a positive of the third separation. The re- 
sultant soft gelatin image was dyed with a pinatype color. These are dyes 
which stain soft gelatin only. The ability of silver iodide to fix dye was utilized 
by W. V. D. Kelley for the formation of gelatin filters. These consisted of 
dispersions of silver iodide plus basic dyes, in gelatin. 

Instead of using silver images from which to form the mordants, R. Gschopf 
(Eng. P. 279381, 311833) used a mixture of ferric oxalate and silver nitrate. 
This was light-sensitive, and upon exposure to light, formed metallic silver. 
The light acted upon the ferric salts to reduce them to ferrous condition. 
These reacted with the silver nitrate to form metallic silver. The image was 
then converted into cuprous thiocyanate by treatment with a mixture of 
cupric thiocyanate and ammonium oxalate. This was dyed with a basic dye. 
The plate or film was resensitized, and the procedure was repeated until three 
dye images were obtained. Instead of mordanting basic dyes, the mordants 
can be made to act upon other constituents, which can then be converted 
into dyes. The silver-iodide complex takes on a negative charge, hence it 
will adsorb other positively charged particles whose sizes are such that true 
crystalloidal solution is improbable. Murray and Spencer (Eng. P. 363616) 
found that diazonium salts will attach themselves to the mordant, much in 
the same manner as basic dyes. The chemistry of the diazonium salts is 
discussed at some length in the chapter on Diazotype Photography. In this 
disclosure, the light-sensitive properties of the diazonium salts are not utilized. 

The silver is converted into a mordant image by any of the well-known 
methods. Thus upon treatment with cupric bromide, it is converted into 
silver and cuprous bromide. If this be treated with 

Potassium iodide 20 parts 

Diazonium salt 2 parts 

Water to 1000 parts 

the bromides become converted into the iodides which then absorb the dia- 
zonium salt. Upon treatment with alkaline solutions of coupling agents, azo 
dyes are formed and precipitated or mordanted directly upon the image. 
Hypo must be present to release the diazonium salt. If the diazonium salt 
is para-diazo-diphenylamine-sulphate, then these can be used as the couplers: 

Magenta 

Borax 5 parts 

Hypo 40 parts 

Potassium hydroxide 0.5 part 

Potassium citrate 1 part 

J Acid 1 part 

% Water to 100 parts 



352 HISTORY OF COLOR PHOTOGRAPHY 



Cyan 




The J Acid is replaced by H Acid. 


^Yellow 




Water 


150 parts 


Potassium hydroxide 


Si parts 


Hypo 


60 parts 


Potassium citrate 


30 parts 


Meta cresol 


saturate 


Water to 


200 parts 



Quinone and naphthoquinone could also be used as mordants. This was 
disclosed by Seyewetz (Brit. J. Phot., Vol. 71 (1924), p. 611; cf. also Ger. P. 
354434). A neutral solution must be used in order to form a mordant. The 
positive is bleached for 10 minutes in the following solution: 



Benzoquinone 


5 parts 


Potassium chloride 


120 parts 


Water to 


1000 parts 



Wash for five minutes, after which the image can be dyed in a 0.1 per cent 
solution of a basic dye containing 0.5 per cent acetic acid. After a five-minute 
wash in water, the image can be cleared with 2 per cent sodium bisulphite. 
The potassium chloride can be replaced by 40 grams of the corresponding 
bromide. The benzoquinone could be replaced with twice the quantity of 
naphthoquinone. The I.G. (Eng. P. 485861) utilized the oxidation product 
of hydroquinone as a mordant. The exposed plate is first developed in a 
mixture of equal parts of the following two solutions: 



Solution A: 




Hydroquinone 


5 parts 


Water 


500 parts 


Solution B: 




Sodium carbonate 


30 parts 


Potassium bromide 


1 part 


Water to 


500 parts 



This yields a silver image plus one consisting of the oxidation products of 
hydroquinone, which is termed a "residual image." The rest of the unexposed, 
hence undeveloped, silver bromide is now exposed, then developed with a 
solution containing a very high concentration of sulphite. Therefore no residual 
image will be formed in this case. Upon converting all the silver to silver 
bromide by treatment with a mild oxidizing agent, the plate will take on a 
uniform sensitivity. If duplitized film is used, there will be present two 
residual images composed of oxidation products of hydroquinone. The third 
image is printed and developed normally. The mordant power of the residual 
images allows this to be dye-toned, after which the silver image could be 
converted into the third color. Basic dyes, of course, are used. f 



DYE TONING 353 

Although this procedure is directly opposite to that first disclosed by Seye- 
wetz, it is quite possible that the same mordant is used in both cases. When 
quinone in the presence of halide reacts with silver, this becomes reduced, not 
to hydroquinone, but rather to quinhydrone. When hydroquinone is oxidized 
in the presence of an excess of hydroquinone, quinhydrone is also the product 
formed. It might be mentioned that the ability of oxidized phenolic developers 
to mordant basic dyes has been known since 1920 when Lumiere and Seyewetz 
published a paper describing this property (Photo. Korr., Vol. 56 (1920), 
p. 140). W. Eller (Ber., Vol. 53 (1920), p. 1469) also disclosed that the 
air oxidation products of pyrocatechin, hydroquinone, or pyrogallic acid will 
tan gelatin, reduce ammoniacal solutions of silver, and form precipitates with 
basic dyes. A complete study of these substances was published by R. Jodl 
(Zeit. wiss. Phot., Vol. 37 (1938), p. 111). 

Dr. Bela Gaspar, who has developed the successful silver-dye-bleach process 
which carries his name, made his first disclosures in the field of mordants 
(U.S.P. 1956122; Eng. P. 369616). Here is disclosed a monopack containing 
two layers, each sensitized to a different portion of the spectrum. After ex- 
posure, development, and fixation, there are obtained two silver deposits, one 
in each layer. These are converted into mordants, then dyed with fuchsin. 
The upper image is decolorized by treatment with nitrous acid, with the 
formation of a diazonium salt. The final step is to couple this with a coupling 
agent to yield a color complementary to that of fuchsin. Evidently diazotized 
fuchsin retains its basic characteristics and remains attached to the mordant 
image. 



CHAPTER 22 

PRIMARY COLOR DEVELOPMENT 



O, 



*NE of the most elegant solutions to the problem of forming a colored 
image, lies in the utilization of the products formed by the action of the de- 
veloper upon the latent image. By this means there is formed a dye image 
whose intensity follows closely that of the silver. This possibility was first 
realized by Dr. B. Homolka not long after the significance of the group rela- 
tions in an organic developer was disclosed by the Lumieres, Seyewetz, and 
Andresen. In order to have a clear and concise view of the matter, a review 
of the fundamentals would not be out of place. 

Photography, as it is practiced today, can be traced back to Talbot and 
the Calotype process, rather than to Daguerre and the Daguerreotype process. 
This was disclosed in 1840. A sheet of paper, sensitized with silver iodide 
plus silver nitrate and gallic acid, after exposure, was washed with a solution 
of silver nitrate and gallic acid. Here was grasped the idea of the develop- 
ment (physical) of a latent image. Several years later, R. Hunt discovered 
that ferrous sulphate could replace the gallic acid. In 1847 Blanquart-Evrard 
found that it was not necessary to give the plate the forebath, but that all 
the steps of development could take place after exposure. In 185 1 Regnault 
and Liebig pointed out that pyrogallic acid could replace gallic. In these 
experiments, acid pyro was used, together with silver nitrate, so that it was 
physical development that was being practiced. In 1862 {Brit. J. Phot., Vol. 9, 
p. 425) Major Russell disclosed the use of alkaline pyro as a chemical developer 
for the latent image. Two stock solutions were prepared, one containing pyro 
and citric acid, the other containing ammonia or some other alkali. The two 
were mixed just before use. Once mixed, the solution had no keeping qualities, 
so that it had to be discarded immediately after use. About 1881, H. B. 
Berkeley discovered that the "addition of sulphite to the pyro solution gave 
the developer keeping qualities. Since then, all developers have been com- 
pounded with much sulphite present, except when this substance interfered 
with certain desired actions. The developing property of hydroquinone was 
discovered by Abney in 1880. In 1887 Eder and Toch disclosed pyrocatechin 
as a developer. The next year Andresen gave us p-phenylene diamine, and 
shortly after that p-amido-phenol. 

It was generally recognized that when a developer acted upon a latent image, 
the developing agent became oxidized. What the photographic chemists did 
not realize was that the oxidation product of the amine or phenol was a dye 

354 



PRIMARY COLOR DEVELOPMENT 355 

capable of staining the gelatin. This realization came relatively late, approxi- 
mately in 19 10, when J. Desalme wrote a paper in which he discussed the 
probable reaction that took place (Ch. Ztg. (1910), p. 953). 

Among other things, he had the following to say: "The oxidation products 
of the organic developers are not quinones, but condensed (polymerized) 
quinone bodies. Thus paraphenylenediamine, by oxidation by air, ferricya- 
nide, or peroxide in alkaline solution, does not yield the quinone-imide 

but yields instead, tetramido-diphenyl-para-azo-phenylene 

(H 2 N)2=C 6 H 3 — N=C 6 H 4 =N— C 6 H3=(NH 2 ) 2 

These oxidation products are, as a rule, dyes that stain gelatin. Because of 
this tendency to stain, these substances could not be used as developers without 
the addition of other compounds, such as sulphite, which would prevent the 
dye formation. By development with a solution containing para-amido-phenol 
and meta-toluylene diamine, there results a blue coloration due to the oxida- 
tion of the amino phenol, and subsequent formation of indaniline. The leuco 
derivatives of the indophenols, indamines, indanilines, and the diphenyl-amine 
dyes in general, can be used as developers. 

It is rather a sad fact that this was not realized earlier, and to make it 
worse, there is no reason why this discovery should not have been made. In 
1856 Perkins succeeded in oxidizing a mixture of aniline and toluidine to make 
the first synthetic dye. Within a decade, this reaction or one similar to it, 
was used photographically. This was in the Willis Aniline Process (Eng. P. 
2800/64). A paper was sensitized with a mixture of dichromate and phos- 
phoric acid. After drying, it was exposed under a copy. Where the light 
acted upon the dichromate, this oxidizing agent became reduced, hence had 
no effect upon aniline vapor with which the paper was fumed. If the acid 
content was low, the formation of a dye image was rather slow, and reddish 
brown tones were obtained. If the acid content was high, much faster action 
resulted, with green or blue tones. Toluidine, pyrrol, or their salts could 
replace the aniline. 1 

Almost simultaneously with this came a disclosure by E. Kopp in the 
Photographic News (Vol. 8 (1864), p. 147). A paper is sensitized with a mix- 
ture of dichromates. After exposure, it is washed free of unexposed salts, 
then treated with aniline, naphthalene derivatives, etc. It is claimed that 
the reduced chromic oxides have associated with themselves chromic acid, 
which oxidizes amines to form dyes. This paper was published also in Chemical 
News (Vol. 9 (1864), p. 296; Vol. 10, p. 28), and in the Photographic News 
about fourteen years later (Vol. 22, (1878), p. 467). 

These and analogous processes were discovered and rediscovered with great 
regularity, but no one appears to have made the rather obvious observation 



356 HISTORY OF COLOR PHOTOGRAPHY 

that the reaction could serve as a basis for color. Among the last to patent 
the idea, were Gusserow and Andresen of the Agfa Company of Germany 
(Ger. P. 116177; Eng. P. 12313/00). The sensitized paper was exposed, then 
washed free of unreacted dichromate. 
It was developed by treatment with: 

Water 600 parts 

Paraphenylenediamine 1 part 

Sodium bisulphite 1-2 parts 

A dark brown image resulted. Other colors were obtained if the p.p.d. was 
replaced by other amines and phenols. Suitable for use were the dimethyl 
derivative of p.p.d., the corresponding toluene derivatives, 1:5 di-amino- 
naphthalene, 1:5 di-oxy-naphthalene, para-amino-phenol, metol, ortol, ami- 
dol, pyrogallic acid, para-diamino-diphenyl-amine, aniline, and di-methyl- 
aniline. This was in 1900. 

The next year they made a real contribution (Eng. P. 5879/01). They dis- 
covered that by mixing several of the substances together, dye images of the 
indamine and indo-phenol class would be formed. This disclosure was the 
final one which fixed completely the chemistry of color development. It is 
immaterial whether the oxidation of the ingredients is accomplished by a 
latent image residing upon silver halide grains, or upon a reduced chromic 
oxide. In either case the same dye is formed. It is a rather curious type of 
astigmatism that must have afflicted Andresen, when he failed to recognize 
that these solutions would be capable of developing a silver image. After all, 
he was the discoverer of the developing properties of the diamines and amino 
phenols, and he had by this time already fully disclosed the structural re- 
quirements for developer action. Gusserow and Andresen seemed to lay con- 
siderable stress upon the fact that it is the reduced dichromates which serve 
as the oxidizing agent. They claimed this to be an original discovery, forgetting 
completely that more than thirty-five years before them, E. Kopp disclosed 
the identical phenomenon. 

In the meantime the practical photographers were also tackling the prob- 
lem. Liesegang noted {Phot. Archiv., Vol. 36 (1895), p. 115; Phot. Chem. 
Stud., Vol. 2 (1895), p. 28) that the color of the image developed with pyro 
was much browner than the one produced by ferrous oxalate. The next year 
A. Watkins, who devised the factorial system of development, proved that 
the brown stain was not silver, and that it was proportional to the image 
density {Phot. J., Vol. 36 (1896), p. 245). He developed an image with a 
sulphite-free pyro developer, then removed the silver by treatment with cupric 
bromide and hypo. Luppo-Cramer proved that the stain was due to an organic 
dye {Phot. Korr., Vol. 42 (1906), p. 319; Vol. 43, p. 242). By this time the 
publications of Andresen {Farbenindtcstrie (1889), p. 187; Phot. Mitt., (1891), 
pp. 124, 286, 296; Jahr. Phot., Vol. 7 (1893), p. 418; Phot. Korr., Vol. 36 
(1899), p. 635; Vol. 37 (1900), p. 185) and of the Lumieres and Seyewetz 



PRIMARY COLOR DEVELOPMENT 357 

(Jahrbuch, Vol. 9 (1896), p. 62; Vol. 12 (1898), p. 100); Vol. 13 (1899), 
p. 306; Vol. 18 (1904), p. 99; Bull. Soc.franq. Phot., II, Vol. 14 (1898), p. 158; 
Brit. J. Phot., Vol. 56 (1909), p. 627), completely outlined the structural chem- 
istry of the developers in the benzene series. It remained for Dr. B. Homolka 
to combine the theoretical findings with the practical {Phot. Korr., Vol. 44 
(1907), p. 55). He recognized that the essential group relationship that must 
exist in developing agents was present in the leuco derivatives of most dyes, 
hence these should act upon a latent image with the subsequent formation 
of a dye. Although he failed to find any leuco bases which gave insoluble 
dye images, he did disclose the developing, properties of indoxyl and thio- 
indoxyl, and the subsequent formation of dye images composed of indigo and 
thioindigo respectively. The solutions he used were the following: 

Blue image 

Sodium sulphite, 6% solution 10 parts 

Potassium bromide 0.5 part 

Indoxyl 1.5-20. parts 

Water to 100 parts 



ed image 




Sodium sulphite, 6% solution 


10 parts 


Sodium hydroxide, normal 


10 parts 


Thio-indoxyl 


1.5 parts 


Water to 


100 parts 



Several years later R. Fischer patented the practical application of Homolka's 
discovery, although instead of using indoxyl and thio-indoxyl he used the 
corresponding carboxylic acids. The — COOH group split off during the re- 
action, yielding the dyes indigo and thio-indigo for the final images. The 
solutions Dr. Fischer suggested were as follows (Eng. P. 15055/12; Ger. P. 
257167; U.S.P. 1055155): 



Yellow image 




Pyrogallic acid 


1 part 


Potassium carbonate 


5 parts 


Water to 


100 parts 


Magenta image 




Thio-indoxyl-carboxylic acid 


0.5 part 


Acetone 


5 parts 


Potassium carbonate 


5 parts 


Water to 


100 parts 


Cyan image 




Indoxyl-carboxylic acid 


0.5 part 


Acetone 


5 parts 


Potassium carbonate 


2 parts 


Water to 


100 parts 



358 HISTORY OF COLOR PHOTOGRAPHY 

In another patent issued simultaneously with the German patent mentioned 
above, and assigned to the Neue Photographische Gesellschaft, Fischer uti- 
lized these developers to form a screen plate. A gelatin-silver-bromide plate 
was exposed under a black-and-white matrix, whose transparent areas were 
one-third the size of the opaque ones. It was next developed with indoxyl, 
thus converting the exposed regions to a blue color. It was then exposed 
again, but this time in a manner which formed a latent image in half of the 
areas previously left unexposed. Development with thio-indoxyl gave a red 
dye deposit. The emulsion was finally exposed completely, forming a latent 
image in the remaining clear areas. Development with chlor-indoxyl gave a 
green image. The silver was removed, leaving only stained gelatin behind, 
upon which could be coated a panchromatic silver bromide emulsion suitable 
for screen plate processes. 

In the same disclosure, Fischer suggested that it would be possible to add 
the agents to the emulsions before coating. The indoxyl could be added to the 
red-sensitive emulsion, etc. These could then be coated upon a plate in 
the form of superimposed layers. It is also possible to tan the emulsions con- 
taining the agents. This would seriously restrict their mobility. The emul- 
sion grains could then be mixed with the probability that the indoxyl would 
remain in the close vicinity of the red-sensitized grains, and not diffuse out 
to the green- or blue-sensitive grains. Upon development, indoxyl will act only 
upon the grains exposed to red light. Since it is proposed to use thio-indoxyl 
to form the primary red required in screen plates, and the magenta required 
in subtractive processes, it can be imagined that it was not suitable for either 
of these purposes. The same is true with indoxyl. Perhaps that is why 
nothing further seems to have been done along these lines. 

The chemistry of developers and the developing agents plays a great role 
in color development. The constituents of an ordinary developing solution 
are the developing agent, sodium sulphite, potassium bromide, and alkali. 
The silver bromide that contains a latent image becomes reduced to metallic 
silver, forming sodium or potassium bromide at the same time. Hence, it 
becomes possible to control the reduction of the silver salts by the concentra- 
tion of bromide ions present in the original solution. The potential of the 
developer will vary inversely with concentration of bromide. Since the action 
upon non-image portions is also dependent upon the potential, the fogging 
tendency of a developer can be controlled by the bromide content. Negative 
developers are compounded with a minimum of bromide present, since here 
it is important to obtain the last ounce of action from the solution. Positive 
developers are compounded with relatively large quantities of bromide present, 
to insure the complete absence of fog. 

Although the action of pyrogallic acid had been known since 1851, it was 
not until approximately thirty years later that it was made stable by the 
addition of sodium sulphite. The brown discoloration that was typical of 
pyro solutions became completely eliminated when sulphite was present. It 



PRIMARY COLOR DEVELOPMENT 359 

was well-known that the brown color was due to the oxidizing action of air 
upon alkaline solutions of pyrogallic acid, hence it was popularly assumed 
that sulphite had greater affinity for air than alkaline pyro. This belief per- 
sisted for some time, but it was attacked by Andresen (Phot. Korr., Vol. 35 
(1898), p. 445; Vol. 36 (1899), pp. 396 and 635; Vol. 37 (1900), p. 185); Bo- 
gisch (Phot. Korr., 37 (1900), pp. 89 and 272); Luther and Leubner (/. prakt. 
Chetn., Series II, Vol. 85 (1912), p. 289; Vol. 86 (1913), p. 41; Vol. 22 (1922), 
p. 72; Vol. 27 (1930), p. 536; Vol. 98 (1918), p. 81); and by Seyewetz and 
Szymson (Bull Soc. Chitn. franq., Series IV, Vol. 53 (1933), p. 1260). Accord- 
ing to Andresen, the sulphite became oxidized, so that the sulphur changed 
from a quadrivalent to a hexavalent state. But it was not sulphate that was 
formed, but rather a sulphonated oxidized developer, thus: 

ON* 



j + 2A^Br^=^| |+2A<2+2NaBr 



oW d 

Di-sodium hydroquinonate Quinone 

o qH 

j + N* 2 S03+H 2 O— >Q. s03Na +NaOH 

S OH 

Quinone Sodium hydroquinone sulphonate 

ONa 9 

0.so^ + ^ B ^0-so iNa +2A ^ 2Na ^ 

o'Na 6 

Tri-sodium hydroquinone sulphonate Sodium quinone sulphonate 

Combining the three steps, the reaction may be written: 
ONa 

+4A^BH-Na 2 SO a -fH 2 0->| |J N +4^+4NaBr+NaOH 



This would indicate that four molecules of silver bromide are utilized by every 
molecule of hydroquinone. The later work of Pinnow and Seyewetz and 
Szymson indicates, however, that the final product of the reaction is not qui- 
none sulphonate, but hydroquinone monosulphonate and disulphonate. These 
can act as developers at the pH of caustic soda solutions, but not at that of 
carbonates. 

When developers are used without sulphite present, the reactivity of the 
oxidized developer makes itself felt in an entirely different manner. In the 



360 HISTORY OF COLOR PHOTOGRAPHY 

absence of any other disturbing substances, polymerization takes place to yield 
brown amorphous residues. A study of this type of developer has been made 
by R. Jodl (Zeit. wiss. Phot., Vol. 37 (1938), p. in). It had previously been 
found that sulphite-free pyrogallic acid, pyrocatechin, and hydroquinone, 
formed substances by their oxidation by means of the latent image, which 
tanned gelatin (Ger. P. 309193, 358093 and 358166, issued to G. Koppmann; 
and 565 1 1 1, issued to F. Leiber). Koppmann in Germany, and the Techni- 
color Motion Picture Corporation in the United States (U.S.P. 1535700; Eng. P. 
204034; Ger. P. 400951; Fr. P. 570076) have made this reaction the bases 
of color processes. The Lumieres and Seyewetz noted that the oxidized de- 
veloper products were able to mordant basic dyes (Phot. Korr., Vol. 56(1921), 
p. 140). This was later made the basis for a color process by the I.G. (cf. 
chapter on Dye Toning). F. Leiber then disclosed (Ger. P. 565111) that it 
was also able to reduce silver nitrate, a property that could be used to in- 
tensify a silver image. 

The air oxidation products had the same properties, a fact that was first 
pointed out by W. Eller (Ber., Vol. 53 (1920), p. 1496). He called the polymers 
formed from phenolic developing agents, huminic acids. They could be pre- 
cipitated from solution by the addition of aniline, quinoline, pyridine, etc., 
with which they formed insoluble salts. They united with gelatin to form 
leather, and with basic dyes to form lakes. They also reduced silver salts 
and Fehling's solution. When treated with diazonium salts, they formed 
azo dyes, all more or less brown in shade. Eller assigned the empirical for- 
mula (CeH^AOx to the huminic acids from pyrocatechin (ortho-dioxy-benzene), 
and hydroquinone (para-dioxy-benzene). It is evident, therefore, that an atom 
of oxygen is absorbed by each molecule of developer. The type of union be- 
tween the molecules has not as yet been established. The potassium and 
barium salts have the empirical formulas (Ci 2 H70 6 K) x , (C^HrOsBaOH)* re- 
spectively. This would indicate that the unit of structure was C12H7O5 • OH. 
It is very probable that this is the substance formed when a developer 
is used which contains no sulphite or other substances that can attach them- 
selves to the extremely reactive position formed within the molecule. It is 
to be recalled that Desalme (cf. above) proposed a similar type of reaction 
for the phenylene diamines. 

Although we know practically nothing concerning the reaction that takes 
place when no sulphite is present, we do know that one or more of the posi- 
tions in the benzene ring become extremely reactive. If any substances are 
present which contain active or labile hydrogens, coupling takes place and 
the oxidized developer ions unite with the "coupler" to form a new substance, 
which in many cases is highly colored and is able to stain the gelatin. Color 
developers, as we know them today, are compounded with practically no sul- 
phite present. Instead, molecular equivalent quantities of "coupling" agents 
are added. These unite with the oxidized developer before that substance 
has a chance to polymerize. 



PRIMARY COLOR DEVELOPMENT 361 

The structural requisites for a derivative of benzene to be a developer, were 
outlined by the researches of the Lumieres and Seyewetz in France and An- 
dresen in Germany. By their very nature, developers must be reducing agents. 
But not all reducing agents are photographic developers. There are many 
inorganic substances which are capable of acting in this capacity, notably 
ferrous citrate, oxalate, fluoride, cuprous ammonium oxalate, hydrosulphite 
(under special conditions), hydrazine, hydrogen peroxide, and hydroxy lamine. 
The last three have great theoretical interest, although little practical value 
since during their action they evolve a gas which causes blisters to form 
within the emulsion layer. They have the structures as follows: 

H 2 N— NH 2 , HO— OH, H 2 N— OH 

Hydrazine Hydrogen peroxide Hydroxylamine 

Here two amino, two hydroxy groups, or one of each, are attached to each 
other. These groups must also be present in the benzene molecule in order 
for it to have developing properties, and their location within the molecule 
must correspond to certain definite rules. The molecule of benzene has the 
structure 



CH 



HC? J CH 




usuallv 
Wrttlen 



0-0 



This arrangement of alternating single and double bonds between the carbon 
atoms of a hydrocarbon, is termed conjugation. In an unsubstituted hydro- 
carbon of this type, the positions within the molecule have no special signifi- 
cance. But consider the case where two of the hydrogens are replaced by other 
groups such as amines or hydroxyl radicals. Three different types of sub- 
stitution are possible, namely: 



0-Q- Or 

x 
Para Meta Ortho 



It is only the ortho and para substituted compounds which can be used as 
developers. 
There are three possible types of tri-substituted derivatives, 

xxx 

J 

z 
1:3-5 1:3-4 1:2:3 



XXX 

■ft 0* 0: 



362 HISTORY OF COLOR PHOTOGRAPHY 

Of these, it is only the 1:3:5 substituted compound that does not have de- 
veloping properties. 

This was as far as the disclosures of the Lumieres, Seyewetz, and Andresen 
went. In 1907 Dr. B. Homolka (cf. above) established that indoxyl and thio- 
indoxyl had developing properties. He also suggested that the leuco deriva- 
tives of di- and tri-phenyl methane dyes had the requisite structures, hence 
should be capable of developing a latent image. Lumiere and Seyewetz tested 
several leuco tri-phenyl methane dyes and failed to notice that they had any 
developing properties. This was in keeping with a study which they made 
upon the effect of a ketonic group upon the developing properties of a sub- 
stance (Brit. J. Phot., Vol. 44 (1897), p. 665). In general, they found that a 
ketone group by itself had no effect, provided an aliphatic radical was present 
on one side of the molecule, and a group with developing characteristics on 
the other. Thus 



r/yc -ch 3 H0 |Y 

v OH 



H °"f I " CH * - ' * ,g ^ V 



develop vigorously. If an unsubstituted phenyl group replace the methyl, 
the compound still has developing properties. The substance 



u A Oh 



OH 
OH 



tri-oxy-benzophenone develops quite vigorously, but not as well as the cor- 
responding aceto-phenone. But if any amino or hydroxyl group is present 
in both of the phenyl groups, the compound loses its developing power. Since 
most of these substances are colored, their application to color development 
is immediate. It does not appear, however, that they were studied from 
this point of view. 

In some later work, Homolka extended the number of odd substances which 
developed a latent image and deposited at the same time insoluble dyes 
{Phot. Korr. y Vol. 51 (1914), p. 256 and 471; Vol. 53 (1916), p. 201; Vol. 56 
(1919), p. 387; Vol. 59 (1922), p. 29). These were oxy-iso-carbostyril, whose 
structure is 



CO 



OH 



and symmetrical di-methoxy-pyrogallol. 



OH 

CH.O-fVOCH3 



H *°lr 



OH 

;ch 



PRIMARY COLOR DEVELOPMENT 363 

This last is very interesting in that it is apparently an exception to the rules 
established above. It would merit some study. 

All the other substances mentioned also appear to be outside the limits 
of the established rules. They all have a common grouping however, of the 
type 

Keto form Enol form 

These exist in a type of isomerism known as keto-enol tautomers. It is in- 
teresting to note that in the enol form the two carbon atoms have the con- 
jugated linkage that is also present in the ring itself. In indoxyl an — (NH) — 
group is attached to the end carbon, and also to the benzene nucleus to form 
a fused indole ring. Now again we have a conjugated system in which ad- 
jacent carbons have a hydroxyl and a substituted amino group present. In 
thio-indoxyl, it is a substituted thiol group (SR) which is attached to the 
end carbon. This would mean that a thiol group, in which the sulphur atom 
is firmly attached and in no danger of splitting off, also can endow a substance 
with developing properties. As a rule such compounds give up the (SH) group 
very readily, forming hydrogen sulphide, which promptly reacts to form silver 
sulphide. In oxy-iso-carbo-styril, a similar situation exists. An (NR) group 
is attached to the end carbon, endowing the molecule with developing prop- 
erties. If this were not sufficient, then we can consider the molecule from 
another point of view. It has two hydroxyl groups separated from each other 
by four atoms, linked together by means of a conjugated chain, hence pre- 
senting a hydroquinone type of grouping. 

There is one other property common to all of these substances. The dyes 
which they form are all well-known. They all consist of two identical groups 
linked together by a double bond. Indoxyl, for instance, yields the dye 
indigo 



2 



o?-o?-^b 



Thio-indoxyl yields the dye thio-indigo 



0?-<tf'H50 



Oxy-iso-carbostyril gives the dye karbindigo 



0?~QJ~~$) 



OH OH 



364 HISTORY OF COLOR PHOTOGRAPHY 

Similarly, symmetrical di-methoxy pyrogallol and para methoxy-alpha naph- 
thol react as follows: 

OCH a OCHd 0C Hi 

Di-methoxy pyrogallol Coemlignone 




OCH3 
Methoxy alpha naphthol 

These reactions all yield well-known products, hence their structures are de- 
termined. In this respect they may furnish a clue as to the nature of the 
polymerization which takes place when the simpler developing agents like 
hydroquinone, metol, pyrogallic acid, pyrocatechin, etc., are used. 

The work of Homolka was generalized by J. D. Kendall (IX* Cong. int. 
Phot. sc. appliq., (1935), p. 252) in a paper read before the 9th International 
Congress at Paris. In the more general form, a substance to have developing 
action must possess a conjugated carbon chain, 

— C=C— C=C— C=C— . 

I I I I I I 
H H H H H H 

Two of the carbons in this chain must contain hydroxyl or amino substi- 
tutions, which must be separated by an even number of carbon atoms, thus 

X-(C=C) -Y 

H H 

The number n can be anything from zero up, X and Y represent amino or 
hydroxyl groups. Thus if n is zero, we have the inorganic series hydrogen 
peroxide, hydrazine, or hydroxylamine. The carbons can be part of a ring 
structure. This structural generalization has recently received striking con- 
firmation when K. Maurer and G. Zapp {Phot. Ind., Vol. 33 (1937) p. 9°)> 
and J. Rzymkowski (Phot. Ind., Vol. 33 (1937), p. 91) disclosed that Vitamin C 
and oxytetronic acid could develop 

O 

II II 

y C— C— OH /C— C— OH 

0< || < II 

>C— C— OH \C— C— OH 

h/ l h/ i 

H— C— OH H 

I 
H 2 — C— OH 

Vitamin C Oxytetronic acid 



PRIMARY COLOR DEVELOPMENT 365 

A partial application of this was patented by the I.G. (Eng. P. 459665; Ger. P. 
646516; Fr. P. 807002). Whenever a ketonic group is adjacent to an amino 
methene group, — (HC — NH 2 ), the compound has developing properties, and 
yields a dye image together with the silver. Such a grouping gives rise to 
the possibility of a keto-enal tautomerism 

— C— C— NH 2 ^± — C= C— 

II I II 

OH OH NH 2 

Keto form Enol form 

and this, as was indicated above, satisfies the structural requirements for de- 
veloping action. Such developers need not always have alkali present to 
energize them. An example in the aliphatic series, which is a weak developer, 
is di-ethyl-amino-acetone 

0=C— CH 3 ^ HO— C— CH 8 
(C 2 H 6 )2N— C— H 2 (GHOsN— C— H 
a better developer is amino-aceto-acetic ester 

O O 

H 3 C— C— CH— C— OC 2 H 6 ^± H 3 C— C=C— C— OC 2 H 5 

II I II 

NH 2 HO NH 2 

Of course the corresponding benzoic or other acid derivative would act in a 
similar capacity. Another possibility is 4-amino-5 pyrazolone 

H 

I 
H— C C— NH 2 ;=i H— C C— NH 2 

II I II II 

N 0=0 N C— OH 

\ / \ / 

N N 

I I 

H H 

Keto form Enol form 

4-amino-5-pyrazolone 

Some typical examples of developer solutions are as follows: 

1. Alpha-amino-aceto-acetic ester 6.0 parts 
Sodium sulphite 20.0 parts 
Sodium carbonate 10.0 parts 
Potassium bromide 0.5 part 
Water to 1000 parts 

2. i-phenyl-3-methyl-4-amino-5-pyrazolone 5.0 parts 
Everything else as in the first example. 



366 HISTORY OF COLOR PHOTOGRAPHY 

3. i-p-chlor-phenyl-3-methyl-4-amino-5-pyrazolone 5.0 parts 

Sodium sulphite 5.0 parts 

Potassium bromide 0.5. part 

Water to 1000 parts 

The last two solutions yield yellow dye images together with the silver. 
Ortho-amino phenol has the structure 

-OH 

L NHz 

therefore it is a good developer, and capable of yielding dye images. It would 
appear that this compound and its substitution products are covered in the 
patent above, but the I.G. made doubly sure by another patent (Eng. P. 
457326), issued a short time before the one discussed above. Yellow to red 
images are obtained. Typical examples are the following: 

Yellow dye image 

4 : 5-di-methyl-2-methyl amino-phenol 1 .0 part 

Sodium hydroxide 0.8 part 

Water to 200 parts 



Orange image 




2-methyl amino-phenol 


1 part 


Potassium carbonate 


10 parts 


Water to 


200 parts 


Red image 




4-chloro-2-methyl amino-phenol 


2 parts 


Potassium carbonate 


15 parts 


Water to 


200 parts 



A further extension was made by the I.G. (Eng. P. 482652), when it disclosed 
that developers like pyrogallic acid, ortho-amino-phenol, etc., could be made 
into primary color developers by substituting a heavy residue into the mole- 
cule. Suitable for this purpose are groups such as diphenyl, stilbene, azoxy- 
benzene, 2 : 3-oxy-naphthoic acid, diarylureas, benzthiazol, etc. 

In 1936 and 1937 Karl and Ludwig Schinzel wrote a series of papers in the 
magazine Das Lichibild (Vols, n and 12). In the first of these articles, the 
chemistry of primary color development is discussed (Vol. n (1936), p. 172; 
Phot. Ind., Vol. 34 (1936), p. 942). The application of these findings has 
been protected by Karl Schinzel in a series of disclosures assigned to the 
Eastman Kodak Company (Eng. P. 498869, 498870, 498871, and 498875). 
They found that hydroquinone, and naphth-hydroquinone can yield color 
images directly if heavy substitutions are placed within the molecule. Lemon- 
yellow images can be obtained if 2 : 5-diaryl-, 2 : 5-di-phenoxy-, 2 : 5-diphenetyl- 
hydroquinone, or 2-phenyl-i:4-dioxy-naphthalene be used as the developing 
agent. 



PRIMARY COLOR DEVELOPMENT 367 

fro-, (fro 




OH 



0CH 3 



Orange or red dyes are obtained if acylamino groups are the substitutions. 
Examples of this are: 



OH 
ft nu 



QH 

H f y v 

QH 



H >> 



L/ OH 

2 : 5-benzoyl amino-hydroquinone 2-benzoyl amino- 1 : 4-dioxy-naphthaIene 

The presence of an azo group in a molecule evidently does not affect the 
developing properties, for the deep blue sodium salt solution of 4-phenyl-azo- 
1 : 2-dioxy-naphthalene can act in the capacity, yielding at the same time a 
lemon-yellow dye image. This substance has the structure: 




N=N< C~~/ > 

and can be formed by coupling benzene diazonium chloride with i!2-dioxy 
naphthalene. 

The generators of the lignone dyes can act as primary color" developers. 
Several instances of these have already been mentioned. The symmetrical 
di-methoxy-pyrogallol is one, and 4-methoxy-alpha-naphthol is another. The 
beta-methoxy-alpha-naphthol gives a purplish red, beta-phenyl-alpha-naphthol 
a purple, and tetra-methoxy-alpha-naphthol a blue. 

Substituted ortho-amino-phenols and naphthols are again mentioned (Eng. 
P. 498871). A yellow image is obtained from 3 : 5-dimethyl-2-amino-phenol. 
The corresponding di-chloro, bromo, and cyano derivatives can also be used. 
Ortho-amino-meta-brom-cresol gives reddish-yellow images, as does ortho- 
amino-para-xylenol. Rose to purple colors can be had using alkoxy ortho- 
amino-phenols, alpha-phenyl-amino-beta naphthol, or beta-phenyl-amino-al- 
pha-naphthol. 

-oh r'YVK.Z/ 






R is an alkyl group 



368 HISTORY OF COLOR PHOTOGRAPHY 

It has been noted above that Homolka called attention to the fact that 
the leuco derivatives of many dyes should be developers. The action of leuco 
indamines, indo-anilines and indo-phenols, was long known. These are all 
derivatives of diphenyl amine 

so that if the ortho or para positions in either of the phenyls be occupied by 
an amino or a hydroxy group, that portion of the molecule would contain the 
configuration necessary for developer action. Some of the more complicated 
leuco di-phenyl-amine dyes are quite stable, since it has been proposed to add 
them to the emulsion before coating (cf. Dieterich, Allison, and Detracolor in 
chapter dealing with Monopacks). Upon oxidation by means of the latent 
image the corresponding dye is formed. Fischer patented the use of these 
substances as color developers (U.S.P. 1102028; Eng. P. 2562/12; Ger. P. 
253335), but, as noted above, Homolka mentioned their use in this respect 
in 1907, and Desalme in 1910 (cf. above). In their articles, and in the 
English patent 498875, the Schinzels turn to the leuco derivatives for color 
developers. They carefully exclude the leuco indophenols, indoanilines, inda- 
mines, and azomethines, as these have been previously disclosed. But they 
do list the leuco thio-indigos, seleno-indigos, etc. In using these substances, 
all air must be excluded. A slight excess of hydrosulphite should be present 
to act as a preservative. 

The fact that the leuco derivatives of the vat and diphenyl methane dyes 
have developing properties, does not mean that all such substances have like 
properties. Thus Seyewetz and the Lumieres were unable to find these among 
the leuco derivatives of the tri-phenyl methane dyes. The azo dyes form 
hydrazo compounds by reduction under controlled conditions, compounds 
which are derivatives of hydrazine. This, as was indicated above, is a de- 
veloper, albeit an impractical one. Phenyl hydrazine also is a developer, a 
fact pointed out by E. Votocek {Brit. J. Phot., Vol. 45 (1898), p. 633). This 
has the structure: 

<( )>N»*H 2 

The hydrazo compounds have the structure: 

■<Z> K <Z> 

If we place an amino group in the ortho or para positions, we obtain: 

H a N<^ )>-M-N <^ ^NHi Or <( )-N-n/ ^> 

In either case, both portions of the molecule have developing properties, and 
upon oxidation by the latent image there should be formed the corresponding 
azo dyes 



PRIMARY COLOR DEVELOPMENT 369 

<0 M -0 »— C>*0 

This reaction opens up the entire field of azo dye chemistry for possible 
application to color development. It may be that the latent image does not 
have a sufficient potential to oxidize the hydrazo compound. In that case, 
perhaps the quinones formed by the action of the latent image upon hydro- 
quinone would serve. These are relatively strong oxidizers, especially in the 
presence of halides, which would serve to remove the silver from solution. 
The hydrazo compounds can be formed by the controlled reduction of azo 
dyes, usually by means of alkaline alcohol and zinc dust, sodium amalgam, 
etc. They can also be prepared by the reduction of nitro compounds with 
zinc dust and alkali. 

H H 

R-N-N-R' Z -^?2 R-A4-R' 
alcohol 

H H 

R _ N 2 Zn + K0H , R-i-i-R 

In the first reaction, unsymmetrical hydrazo compounds are formed. In the 
second, symmetrical compounds are formed. Another possibility lies in the 
use of the leuco di- and tri-phenyl methanes, together with hydroquinone. 
The quinones that are formed should oxidize the leuco derivative to the dye. 
If this is insufficient, then it is possible to convert the silver into a substance 
that is capable of oxidizing action. The disclosures of Willis and Kopp who 
oxidized aniline and other organic substances by dichromate or dichromate 
reduction products, have already been discussed. The production of dyes by 
such a procedure was suggested by Gusserow and Andresen (cf. above). These 
experimenters did not utilize a silver image, but one composed of chromic 
and chromous oxides formed by the action of light on dichromate sensitized 
emulsions. The Neue Photographische Gesellschaft in a series of disclosures 
(Eng. P. 18370/03, 10898/04, and 21584/06; Ger. P. 157411, 180947, and 
180948) showed how a silver image could be converted into one of manganese 
oxide. This could then be made to react with amines or phenols to form 
colored residues. The silver image is first treated with one of the following: 

A. Potassium ferricyanide, 0.5% solution 100 parts 
Manganous sulphate, 2 % solution 20 parts 
Potassium bromide, 10% solution 15 parts 

B. Potassium ferricyanide, 0.5% solution 100 parts 
Manganous sulphate, 2% solution 25 parts 



370 HISTORY OF COLOR PHOTOGRAPHY 

C. Manganous ferricyanide, saturated solution 200 parts 
Potassium bromide, 10% solution 10 parts 

After bleaching the silver, the print is washed and treated with: 

Potassium ferricyanide, 2 % solution 9 parts 

Sodium hydroxide, 4% solution 10 parts 

In this solution an image composed of manganese oxide is formed. This is 
finally treated with the color formers. 

The catatype process of Ostwald and Gros (Jahrbuck, Vol. 17 (1903), p. 519) 
operated along a similar idea. A silver image is toned so that it becomes 
converted into one of platinum. If this be brought in contact with a paper 
that is sensitized with pyrogallic acid and potassium bromate, the pyro will 
be oxidized in contact with, and to an extent proportioned to, the image 
density, forming a stain. Copper sulphate is also useful as a catalyst. If an 
image of silver or platinum is flooded with peroxide, this substance will be 
decomposed in situ with the metal. The other portions will remain intact. 
By this means a peroxide negative image is left. The peroxide can then be 
made to oxidize amines or phenols to form dyes. 

E. R. Bullock bleached the silver image with a mixture of ferricyanide and 
chromic acid or permanganate (U.S.P. 1279248). By this means there was 
formed an image of silver ferricyanide. The reduced manganous salts could be 
removed by washing with a five per cent oxalic acid solution. Great care must 
be taken that no halides are present as otherwise silver halide will be formed. 
After a thorough wash, the bleached image is treated with amines. Benzidine 
gave a strong blue, ortho-tolidine or dianisidine gave a green, and paraphenyl- 
enediamine or alpha naphthylamine, a purple. If the bleached image be 
treated with manganous bromide, silver bromide and manganous ferricyanide 
would be formed, making the subsequent oxidation of the amines easier to 
carry out. The silver can be fixed out with hypo to give greater brilliance. 

A similar idea is utilized by Dr. Bela Gaspar (U.S.P. 1956017; Eng. P. 
379679). The esters, ester salts or ethers of the leuco bases of dyes are in- 
corporated into an emulsion. These are not affected by the operations of ex- 
posure, development, fixation, or wash. The silver image is next converted 
into a substance that is capable of regenerating the dye. A photographic 
emulsion is prepared to contain diacetyl indigo white. After exposure, de- 
velopment, and fixation, the silver is converted by well known means into 
lead chromate. As long as no free chromic acid is formed, no action will take 
place with the leuco dye. To accomplish the formation of lead chromate, 
the image is first treated with lead salts plus potassium ferricyanide in the 
presence of ions like citrates, lactates, etc. The lead ferrocyanide is then 
treated with neutral dichromates or chromates. When an image composed 
of lead chromate is acidified, chromic acid is formed, which reacts with the 
leuco dye to yield the insoluble blue dye, indigo. The unreacted leuco dye 
as well as the lead can be removed by treatment with fixing agents and acetone. 



PRIMARY COLOR DEVELOPMENT 371 

A red color can be formed if indigo white be replaced with the sulphuric ester 
of the leuco compound 6:6' di-brom-di-methyl-bis-thio-naphthene-indigo. 
Yellow is formed from an ester of the leuco derivative of helindon yellow. 
The fact that substances such as 4-methoxy-alpha-naphthol, and 2:6-di- 
methoxy-phenol 




' x OCH 3 

6ch 3 




are developers, yielding dyes whose structures are the following: 

3H 9 9 

z ' 





WHs OCH3 OCH3 

and 
QCH3 0£H* QCH3 

2hoO->°=0=0>= 

0'CH 3 OCH3 0'CH 3 

gives rise to the possibility that it may be possible to make other developers 
that cannot form quinone imide structures by oxidation, but form, rather, an 
indigoid type of structure. Since para methoxy-phenol is not a developer, 
whereas the di-methoxy compound is; and since alpha naphthol is not a de- 
veloper, but para-methoxy-alpha-naphthol is; the inference may be drawn 
that in heavy molecules the structural requirements for the presence of de- 
veloper power are different from what they are in the simpler compounds. 
It is well known, for instance, that phloroglucinol 

H<X/V0H 



HO 

cannot act as a developer, but that tri-methyl-phloroglucinol 



H30A 



1CH3 



H0\/OH 
CH 3 

is a developer. Also it is well established that resorcin does not reduce ex- 
posed silver halide grains, but tri-methyl resorcin 

OH 

H3CA-CH3 
IJoH 

CH3 



372 HISTORY OF COLOR PHOTOGRAPHY 

can develop the latent image. In these compounds there is no possibility for 
the formation of ortho- or para-quinoid structures, nor is there the possibility 
that the carbon ortho or para to a hydroxyl group can become highly reactive 
in the same manner as in methoxy-alpha-naphthol. It would be extremely 
interesting to determine whether symmetrical" di-methoxy-phenol 




OH 
cr 



H3C-ACH3 



OCH 3 HjCOV 

will develop, and if so, whether dye formation does not also take place simul- 
taneously. 
Another possibility lies in the compounds of the type 

Hb-kZDH ' «• [(ch j ) / n<3- oh ] ci 

Here there is destroyed any likelihood that para-quinoid structures 

can be formed. But that does not preclude the absence of developer action, 
since there always remains the possibility that two oxidized residues will couple 
to yield a quinoid structure such as: 



2[(CH3) 5 N<^3-NH,]CI 





CH-/-CH3 N C CH 3 
S, cU 3 Otis tH 3 

It can reasonably be supposed that the equivalent naphthalene compound 
would act in this manner, since we have the evidence with methoxy-alpha- 
naphthol. The methoxy group is strictly a neutral one in so far as giving 
a compound developing power is concerned. The quaternary ammonium 
nitrogen may not be so neutral. 



CHAPTER 23 
COUPLING COLOR DEVELOPMENT 



I 



N the preceding chapter we discussed primary color development, in which 
the dye image is formed by the direct oxidation of the developer itself. Al- 
though this method is very simple and most direct, and although it was in- 
troduced some time before all other methods, it is not the technique which 
is being favored at the present moment. The fact that an oxidized phenylene- 
diamine can unite with other amines and phenols was known for a long time, 
for the indophenol, indamine and indaniline dyes were prepared in that man- 
ner. But their application to color photography did not come until 1901 when 
Gusserow and Andresen amplified their disclosure of the previous year. In 
the first patent (Eng. P. 12313/00; Ger. P. 116177) they disclosed that 
amines and phenols become oxidized to colored substances by the product of 
the reaction between light and dichromates. The list of chemicals which 
could be used were as follows: 

Para-phenylene-diamine Amidol 

Di-methyl-para-phenylene diamine Pyrogallic acid 

Para-toluylene-diamine 1 : 5 Dioxy-naphthalene 

1 : 5-Di-amino-phenol Para-diamino-diphenyl-amine 

Para-methyl-amino-phenol Aniline 

Ortol Di-methyl-aniline 

The paper, sensitized with a strong solution of ammonium dichromate, is 
exposed under a negative. It is washed free of unreacted dichromate, then 
treated with : 

Amine or phenol 1 part 

Sodium bisulphite 1-2 parts 

Water to 600 parts 

The different substances give different colors. In this form, the procedure 
offers no novelty over the one proposed by E. Kopp in 1864 (cf. preceding 
chapter). But in the next disclosure (Eng. P. 5879/01; Ger. P. 123292) a 
real advance was made. Instead of having present but one of the agents, 
several were used, one of which was paraphenylenediamine or an aminophenol. 
Now upon treatment with the exposed and washed plate, a dye image of the 
indophenol, indamine or indaniline was formed. 

* In 1908 Henri-Raymond Vidal treated paraphenylenediamine with dichro- 
mate and obtained a color image (Fr. P. 391465). He, as well as Gusserow 

373 



374 HISTORY OF COLOR PHOTOGRAPHY 

and Andresen, failed to notice that these solutions could be used as developers 
until after Fischer made his disclosures and together with Siegrist published 
a full account of the chemistry and properties of such solutions. In 1914, 
a year after the fact was made known generally, Vidal made the discovery 
that his solutions also had developing power (Fr. P. 468537). Instead of 
using pure dichromate as the sensitive medium, A. Thiebau (La PkoL (1908), 
p. 227; Brit. J. Phot., Vol. 55 (1908), p. 738), suggested the following as 
being more sensitive 



Ammonium dichromate 


90 parts 


Copper sulphate 


45 parts 


Manganese sulphate 


10 parts 


Water to 


1000 parts 



After exposure, the paper was washed, then immersed in a 5 per cent solution 
of one of the following agents, slightly acidulated with sulphuric acid. 

Aniline hydrochloride 

Phenols 

Pyrogallic acid 

Cresols 

Naphthols 

Para-amino-phenol 

Para-phenylene-diamine 

etc. 

In 1 91 2 Dr. Rudolph Fischer disclosed the fact that the latent image in a 
silver halide emulsion could replace the reduced chromium oxide images 
utilized by Kopp, Gusserow and Andresen, Vidal, Thiebau, etc. Whether 
or not he was familiar with the work of these. people is a question open to 
debate, for neither in his basic patents (Ger. P. 253335; En g- p - 2562/13, 
5602/13; U.S.P. 1079756 and 1 102028) nor in the article which he wrote 
together with Siegrist (Phot. Korr., Vol. 51 (1914), p. 18) does he mention it. 
But he does discuss the disclosures of Homolka on primary color development 
(cf. preceding chapter). The article covers the fundamentals so thoroughly 
that it well merits the detailed abstraction which follows. It is entitled "The 
Preparation of Dyes by Oxidation by Means of the Latent Image." 

This subject has received but scant attention because it was very difficult 
to isolate and purify the dye formed. But the phenomenon is well known. 
The brown stain of a pyro-developed image is due to an oxidation product, 
and this allows itself to be isolated from the silver. In 1907 Homolka showed 
that indoxyl and thio-indoxyl would develop a latent image and yield the 
dyes indigo and thio-indigo together with the silver. This brings to mind 
the agents para-amino-phenol, and para-phenylene-diamine, which upon oxi- 
dation in the presence of amines and phenols, yield the dyes of the indamine,. 
indo-phenol, and indaniline classes. The leuco derivatives of these dyes have 
already been proposed as developers, but their use has been restricted to the 



COUPLING COLOR DEVELOPMENT 375 

preparation of silver images, completely disregarding the fact that insoluble 
dyes are formed also. This was probably due to the fact that the solutions 
were compounded with large quantities of sulphite present to act as pre- 
servatives, a fact which would completely prevent the dye formation. Here 
is disclosed their use to yield dye images rather than silver. The fundamental 
reactions are as follows: 

HiN<( ^-MJ^ ^>-OH + O z »H 2 N<^ ^>-N=<f = )>0-f2H 2 O 

Para-phenylene-diamine Phenol Indoaniline 

H a N <( )>-N H 2 -f^ y- H U 2 +0 X >H£\<^ )>-^<(3)>»NH-f^H 2 

Para-phenylene-diamine Aniline Indamine 

H0<( )>-N W<(~y H+ Oi >H0<( )>-N^<( = ^>0+2H 2 Q 

Para-amino-phenol Phenol Indophenol 

Instead of oxygen as the oxidizing agent, the exposed silver bromide could be 
used. 

H Z N<( )>NH t 4A^6r-f-<( )>0H >4^4HBHH 2 N<( ^-N=<^ == )>=Q 

HJnj< >NH z -f4A$Brf< >NK 2 > 4AgHHBr+Hn/ >-N»<f C= )>-NH 

H0<( ^>NH a 44A^Br4<^ ^>OH > 4 Agt4H6r+H0<( ^-ffr^"^- 

Since hydrobromic acid is formed during the reaction, it will proceed best 
in the presence of alkali, which will neutralize the acid as fast as it is formed. 
The amino group in para-amino-phenol, and one of them in the diamine, must 
be free of substitutions. The corresponding naphthalene derivatives could 
also be used. Thus instead of para-phenylene-diamine, 1:4 hydrobromic- 
naphthalene could serve, and instead of phenol or aniline the corresponding 
naphthylamine or naphthol could be substituted. 

Fischer and Siegrist gave the name Color Development to this process and 
they termed the second agent present, couplers. Thus phenol is the coupling 
agent in the reaction where indophenols or indoanilines are formed, and aniline 
is the coupler in the third reaction. By reducing the silver halide grain, the 
developing agents are oxidized probably to form quinoid compounds with the 
hypothetical structures 

HN»<^> = N-OH ^ H 2 N-< >-N-0 



376 HISTORY OF COLOR PHOTOGRAPHY 

The form on the left is a hydroxyl amine derivative, while its tautomer on 
the right is a nitroso compound.' These substances could couple or condense 
with compounds that have reactive hydrogens. In aromatic chemistry it is 
well known that the hydrogens in the ortho or para position to an hydroxyl, 
amino, or substituted amino group, are very reactive. We can conceive the 
reaction to be the following: 

Here X represents an hydroxyl, amino or substituted amino group. This 
scheme sounds quite plausible when it is recalled that methods of formation 
of these dyes have been proposed in which nitroso-phenols or amines are used, 

H0-<( )>-N-0 

These substances can be very readily prepared by the action of nitrous acid 
HO — N=0, upon di-ethyl aniline or phenol 

(C 2 H 5 ) 2 N-< )jH + H0;-N=O— ^(Hfe^ N-<C~>*N*0 
H0<( yH+ HOJ-N'O — >H 2 (HH0-<( )>-N»Q 

In order to form this hypothetical compound from the diamine or amino- 
phenol, two atoms of oxygen must be used up, thus 

M<( ^>NH 2 40 2 — > HN-<^>-N-OH + H20 

H0< >NH 2 -K) 2 — > 0*<^>=N-OH+H 2 

This means that a change of four valence units takes place. When silver 
bromide is reduced to silver, with the liberation of bromide ions, a change 
of only one valence unit occurs. Therefore four molecules of silver bromide 
must be reduced in order to supply the requisite electrons for a four-fold 
change in valences. This is discussed in some detail by A. G. Tull {Brit. J. 
Phot., Vol. 85 (1938), p. 627). 

The aromatic amines and phenols are not the only substances that can 
be used as couplers. Any compound which has a reactive methine group, 
— CH 2 — , can be used. Such a group is formed when in methane two of the 
hydrogens are replaced by strongly polar groups, to form R — CH 2 — R', R 
and R' being groups that have strong electric charges. Such groups are phenyl, 
cyano, carbonyl, acetyl, etc. Thus a substance such as phenyl-aceto-nitrile, 
\ y — CH 2 — CN, contains a very reactive — CH 2 — group. Other examples 
are the aceto-acetic-esters, 

CHs— C— CH 2 — C— OR, 

II II 

O 




COUPLING COLOR DEVELOPMENT 377 

aceto-acetanilides, 



CH 3 



With such compounds azo methine dyes are formed. 

Not only did Fischer and Siegrist outline the fundamental chemistry of 
the reaction involved, but they also discussed rather fully the effect of sub- 
stitutions within the molecules used, either for developing or for coupling 
They left practically nothing to be discovered, except tricky and complicated 
substitutions within the molecules, in places where no effect would be had 
upon the chemistry of the reaction. These merely modified the chemical and 
physical properties of the dyes formed, made their spectral absorptions cor- 
respond more closely to theoretical requirements, made them more insoluble, 
and substantive to the gelatin medium in which they were used, etc. No 
new chemical configurations were developed that would couple with the oxid- 
ized developer. 

The substitution of a methyl or ethyl group into the developer molecule 
deepens the color. In a series of tests in which the same coupler, di-chlor- 
alpha-naphthol, was used, and the developing agents varied, the results were 
as follows: 

Developing Agent Color of Dye Image 

Para-phenylene-diamine Pinkish or purplish hue 

Para-toluylene-diamine Blue 

Para-di-methylamino-aniline Cyan 

Para-di-ethylamino-aniline Cyan 

The same type of color change results when the coupler is varied. This is 
indicated in the following series of experiments, in which para-di-ethyl-amino- 
aniline was used as the developing agent. 

Coupler Color of the Image 

Phenol Blue with a greenish tint 

Ortho-cresol Cyan 

Meta-cresol Cyan 

1:4:5 Xy lenol Cyan 

Alpha-naphthol Blue 

Di-chlor-alpha-naphthol Cyan 

Tri-chlor-alpha-naphthol Greener cyan 

Penta-chlor-alpha-naphthol Green 

These give only blue and blue-green shades. In order to obtain yellows and 
red, it is necessary to turn to the active methine group and the azine dyes. 
The same situation holds with these substances that was true with the aromatic 
amines and phenols. 



378 



HISTORY OF COLOR PHOTOGRAPHY 



Using para-nitro-phenyl-aceto-nitrile, O2N — < ^ ) > — CH2 — CN as the cou- 
pling agents, the following variation in colors was obtained by the substitution 
of methyl and ethyl groups within the molecule of the developing agent. 



Developing Agent 
Para-phenylene diamine 
Para-toluylene diamine 
Para-ethyl-ammo-aniline 
Para-di-ethyl-amino-aniline 



Color of the Dye Image 
Orange yellow 
Reddish yellow 
Bluish red 
Magenta 



Using the same developing agent, the color of the dye image in the aceto-acetic 
ester group also deepens. 



Coupler 
CHs— C— CH 2 — C— OC 2 H B 

II II 

o o 

Ace to-acetic-ester 

O-C-CHa-C-OCzHs 
6 6 

Benzoyl-acetic-ester 

H 2 N— CH 2 — C— CH 2 — C— OC 2 H 6 

II II 

o 

Alpha-amino-acetic-ester 

Cyan-aceto-phenone 

NC— CH 2 — CN 
Methylene-cyanide 



Color of the Dye Image 
Yellow 
» 
Reddish yellow 

Orange 

Red 



Magenta 



If the methine group, — CH 2 — , is part of a ring structure, the color is also 
much deeper. 



Coupling Agent 


Color of the Dye Image 


^>-C-CH a -C-0CxH S 


Reddish yellow 


B enzdy 1-acetic-es ter 




OS 



Diketo-hydrinden 


Blue 



COUPLING COLOR DEVELOPMENT 379 



a 



y CH* 

% 

Coumaranon 



' Red 



a 



' ""GHz Magenta 



Thioindoxyl 

The azo-methine dyes have much more stability toward light and acids 
than the indophenols, indoanilines, and indamines. Strong acids decompose 
the dyes into the respective quinone or quinone-imides and the couplers. 
Therefore under no circumstances should strong acids be allowed to come 
in contact with the dye images. Acid short stops and acid fixing baths of 
the conventional type should not be used. Being well aware of these diffi- 
culties, Fischer in his disclosures suggests other and more convenient working 
methods. The three-color separations are first printed as ordinary black-and- 
white prints, on some stripping material. The silver of the images is next 
converted into silver ferrocyanide by treatment with potassium ferricyanide. 
This last could then be treated with coupling developers. The troubles that 
might arise from acid short stops and acid fixing baths are completely elimi- 
nated, since at this stage of the process no dye images yet exist. 

Fischer and Siegrist commercialized their findings by marketing a self-toning 
paper which contained the coupling agents as an integral part of the emul- 
sion. By developing a paper with paraphenylenediamine, a toned print was 
obtained immediately. In this work, they anticipated by £ full generation 
the later work of the I.G. whose Agfa Color contains emulsions with the 
couplers present. The "earlier efforts failed mainly because the appreciable 
solubility of the coupling agents caused them to diffuse out of the emulsion 
in sufficient quantity to affect the developer characteristics. Thus a developer 
became contaminated immediately after use. The I.G. remedied this defect 
by placing substitutions within the coupling molecule, which made that an 
extremely heavy and immobile compound despite relative high solubility in 
alkaline solutions. Diffusion out of a gelatin layer was extremely slow. To 
all intents and purposes, the coupling agent behaved as a colorless "dye" 
that was mordanted to or was substantive to the gelatin. The importance 
which the I.G. attached to this particular phase of the subject may be gathered 
from the fact that no fewer than fifteen United States, and thirteen Eng- 
lish, patents cover it. In general, a very heavy molecular residue is attached 
to the coupler in a position where it will not interfere with the coupling qual- 
ities. Strictly speaking, from a chemical point of view, this procedure is 
merely like substituting an ethyl group for a methyl, within a compound 
which does not depend upon either of these groups for a determination of its 



380 HISTORY OF COLOR PHOTOGRAPHY 

chemical properties. From a patent and legal point of view, the justification 
is that a desired modification in the physical properties is achieved, this 
modification and its importance being fully described in the patent specifica- 
tions. Since these disclosures were discussed at some length in Chapter u, 
they need only be mentioned here. 

A polyvinyl maleic acid residue is attached to a coupling molecule by means 
of an amino group present within that molecule. Thus i-para-amino-phenyl- 
3-methyl-5 pyrazolone is made to react with a polyvinyl-maleic acid anhydride 
to form 

(CH£CH-CH-CH)n ^ 



0=C C-N-/ S-Ni C 'ft H2 
i-(polyvinyl-maleic-anilido)-3-methyl-5-pyrazolone 



This is disclosed in United States patent 2179234 (Eng. P. 468894; Fr. P« 
811541). 

Other heavy groups that can be used are stilbene, azoxybenzene, diphenyl, 
oxy-naphthoic acid amide, benzthiazol, amino phenols and phenols substituted 
in the 3:5 position (U.S.P. 2179238; Eng. P. 458400; Fr. P. 803566). The 
residue can consist of a hydrocarbon with a minimum of five carbon atoms 
in the chain. This can be linked to the molecule by a homopolar bond (U.S.P. 
2186849; Eng. P. 465823; Fr. P. 810410), through an acid amide radical 
(U.S.P. 2186735; Eng. P. 489164; Fr. P. 828603), or by a straight substitu- 
tion for one of the hydrogens attached to a carbon (U.S.P. 218673 1; Eng. P. 
488048; Fr. P. 822166). Resins (U.S.P. 2186733; Eng. P. 489274; Fr. P. 
827625), carbohydrates (U.S.P. 2186732; Eng. P. 483000; Fr. P. 824878), 
polypeptides such as gelatin, albumen, or their degradation products (U.S.P. 
2179244; Eng. P. 484098), sterol or the bile acids (U.S.P. 2186851; Eng. P. 
489093; Fr. P. 827626), polymeric carboxylic acids such as polyglycuronic 
acid or polyacrylic acid (U.S.P. 2178612; Eng. P. 479838; Fr. P. 807792), 
and hydrogenated aromatic or heterocyclic compounds (U.S.P. 2 1867 19; 
Eng. P. 491958) are some of the other substitutions that have been suggested. 
Very useful in this respect is the molecule i-hydroxy-2-naphthoic acid, which 
can be made to react with aromatic amines such as dianisidene to yield cyan 
couplers (U.S.P. 2179228). Two or more molecular residues, each of which 
is capable of coupling, can be condensed to form a single molecule (U.S.P. 
2186734; Eng. P. 489161; Fr. P. 828579). 

The first world war stopped the further development of the subject. In the 
early nineteen-twenties, the Schinzels turned their attention to color develop- 
ment, but they apparently waited until 1936 before they made public any 
of their findings (Lichtbild, Vol. 12 (1936), p. 19). They pointed out that 
Fischer and Siegrist had proceeded under the mistaken notion that an appre- 
ciable solubility was necessary in order for the reaction to take place. But 
they found that this was not so. For instance leuco naphthol blue is ex- 



COUPLING COLOR DEVELOPMENT 381 

tremely insoluble in alkaline solutions. Yet an emulsion prepared with this 
substance present, gave a good image when it was merely bathed in soda after 
exposure. They therefore centered their attention upon the building up of 
heavy coupling agents whose solubility in developer alkalinity would be very 
slight. To accomplish this, they adopted approximately the same dodge used 
by Ansco and the I.G., substitutions within the coupling molecule that made 
it extremely heavy. Very useful were amino groups that could be converted 
into an acid amide by treatment with a compound containing a free COOH 
radical, or vice versa. It must be borne in mind, that whereas the LG. chem- 
ists sought high solubility coupled with extreme immobility, the Schinzels 
strove for extreme insolubility. 

Color development of the secondary type did not become popular until the 
Eastman Kodak Company adopted it as a processing technique for monopack 
film. The first patents issued to Mannes and Godowsky of this company 
which mentioned color development, were in 1932 (Eng. P. 376794, 37^795* 
and 376838). The corresponding United States patent (1954452) was not 
issued until two years later. Dr. M. W. Seymour, another member of the 
Eastman staff had also interested himself in this procedure about this time 
(U.S.P. 1900869 and 1900870), as did Capstaff (U.S.P. 1954346 and 1969452) 
and Burwell (U.S.P. 1966330). These all disclosed various monopack schemes 
(cf . chapter on Monopacks) and contributed nothing not previously discussed 
by Fischer and Siegrist. 

The first patent dealing with couplers that was issued to the Eastman 
Kodak Company, came in 1934, and was issued to Dr. M. W. Seymour (U.S.P. 
1969479). This disclosed the pyrazolones as possibilities for the formation of 
magenta images. The pyrazolones could be prepared by the action of beta 
ketone esters upon hydrazines. Thus when aceto-acetic-ester reacts with 
phenyl-hydrazine, the reaction that takes place is as follows: 

C 6 H 5 -N-!h C 2 Hs04-C* 

CH 3 
Phenyl hydrazine Ethyl-aceto-acetic-ester 

C6Hs-N— ,<° 

cVb 

i-phenyl-3-methyl-5-pyrazolone 

When prepared in this manner, the compound is characterized by the pres- 
ence of a methine group between two polar carbon atoms, hence the hydrogens 
attached to this are extremely active. Not only will they couple with oxidized 
paraphenylenediamine, and amino-phenols, but also with diazonium^salts. 



382 HISTORY OF COLOR PHOTOGRAPHY 

Dr. Seymour gave several formulas which could be used, of which the fol- 
lowing is a typical example. 
Solution A . 

Para-di-methyl-amino-aniline-monohydrochloride 0.5 part 
Sodium sulphite 7.5 parts 

Water to 250 parts 

Solution B. 

i-Phenyl-3-methyl-5-pyrazolone 0.5 part 

Acetone 10 parts 

Add B to A. 

Several years later, the Eastman company was granted several more general 
patents covering the pyrazolones. In the first of these (Eng. P. 478990) a 
heterocyclic group could be attached to the molecule in the number-one posi- 
tion. In the number- three position, furyl, quinolyl, benzthiazolyl and like 
groups, or hydroxy and carboxy groups can be present. The ketone structure 
could be replaced by the equivalent thio-ketone or carbimide radical. In a 
patent issued still later (Eng. P. 496196) there is disclosed the use of two 
pyrazolone molecules linked para to each other through a phenyl group. 

6 

The phenyl groups can be replaced by amino-phenols, -cresols, -naphthols, 
and their alkyl substitution products. The starting point for the compounds 
of this family is the substance 

R0-C-H 2 C-C-/ \ -C-CH 2 -COOR 
o o N — ' o 
Terephthaloyl-bis-acetic-ester 

This upon condensation with two molecules of phenyl-hydrazine, yields para- 
di-( i-phenyl-5-pyrazolone) benzene. 

Not to be outdone, the LG. prepared a pyrazolone derivative starting with 
di-ethyl-malonate (Eng. P. 502665). This has the structure 

0=C— OC 2 H 5 

I 
H— C— H 

I 
0=C— OC2H5 

When this is treated with stearoyl chloride, there is formed : 

0=C— OC2H5 

• HC— Cf 

I X OC 2 H 5 
C^=C — C 17H36 



COUPLING COLOR DEVELOPMENT 383 

In this configuration the molecule was rearranged slightly to put a stearoyl 
radical along the straight chain. Upon treatment with phenyl-hydrazine-3- 
carboxylic acid, a pyrazolone is formed: 

<Zy~f £-&%?£ >H 2 O+C 2 H 5 0H . <Z>-N-CC° H 

Upon heating with acids, ethyl alcohol and CO2 split off, in accordance with 
a well-known reaction, leaving i-(3'-carboxy-phenyl)-3-heptadecyl-5-pyra- 
zolone 

I — / J ^ CH * 
c6oh N^c 

G n H* 

These substances yield red to magenta dye images. The molecule, incidentally, 
is rather interesting. Coupling properties are due to the CH2 group within 
the pyrazolone molecule. Solubility in alkali is attained since an acid group, 
COOH, is present. This is attached to the phenyl group, which plays no part 
in the chemical activity of the molecule. Very low mobility and diffusion 
through gelatin are achieved by the presence of a very heavy group, C17H36, 
and by a very high molecular weight, 443. A molecule as heavy as this cannot 
exist in solution except as a colloid, whose diffusion into and out of gelatin is 
very, very limited. 

One bad feature of color development is that the solutions compounded 
with a minimum of sulphite present have no keeping quality. To improve 
this, E. E. Jelley, of the Kodak, Ltd. staff has suggested the use of sul- 
phoxylates (Eng. P. 462140). These are compounds formed by the action of 
formaldehyde upon sulphites and hydrosulphites. The active group becomes 
masked by union with the formaldehyde. Rongalite C is such a union of 
formaldehyde and sodium hydrosulphite. The sulphoxylates will preserve the 
developer somewhat against air oxidation, but will not interfere to any extent 
with the reaction between oxidized developer and the coupling agent. A 
typical example is the following: 



Solution A. 




Para-diethyl-amino-aniline hydrochloride 


3.0 parts 


Sulphoxylate 


5.0 parts 


Sodium carbonate 


50 parts 


Potassium thiocyanate 


0.5 part 


Water to 


1000 parts 


Solution B. 




Ortho-hydroxy-diphenyl 


2.5 parts 


Acetone 


100 parts 



384 HISTORY OF COLOR PHOTOGRAPHY 

Add solution B to A just before use. A cyan dye image is obtained with this 
solution. If the hydroxy-diphenyl be replaced with naphthoyl-acetonitrile, 
a magenta color would be obtained. 

It is to be noticed that the stock solution A contains potassium thiocyanate. 
This is a silver solvent, whose addition to developer solutions is desirable in 
reversal processes (cf. chapter on Processing Screen Plates). More recently 
it began to appear as an important addition to fine-grain developers such as 
DK-20. The substance was put to a third use by Mannes and Godowsky 
(U.S.P. 2191713) who found that its presence gave a developer solution a 
very pronounced boost in potential. 

The disclosures of Fischer and Siegrist have described rather completely 
the chemical requisites for a substance to act as a coupler. So fully did they 
cover the field that, with the possible exception of pyrazolone and its deriva- 
tives, no new chemical configurations have since been disclosed, but merely 
substitutions made in the old that left the reactive portion of the molecule 
intact. Even in the case of pyrazolone there is present a methine group, 
— (CH2) — , which has the other two valences satisfied by strongly polar link- 
ages, such as a carbonyl (0=C), on one side, and a carbimide on the other, 
( — N=C). This, as all chemists can immediately foresee, makes the methine 
hydrogens very reactive and replaceable. To make matters still more com- 
plicated, the original papers amplified the chemistry of the process and dis- 
cussed the effect which substitutions would have upon the shade of the re- 
sultant dye image. 

The patent office does not as a rule grant protection where the only change 
made is the substitution of an ethyl group for a methyl, since this only modifies 
the physical properties of the compound, without affecting the chemical. But 
this substitution of an ethyl for a methyl group (using these terms now in a 
generic sense rather than in a literal) can be made slightly complicated. For 
instance, Mannes and Godowsky disclosed that hydroxy-diphenyls could be 
used as coupling agents (U.S.P. 2039730; Eng. P. 458665). In the original 
Fischer patent phenol, cresol, xylenol, and naphthol are mentioned. These 
have the constitutions 

QH 

0-OH /\OH /VOH 
\)wi \Ah 3 
ch 3 

Phenol Cresol Xylenol Naphthol 

Any hydroxy aromatic compound that has the para or ortho positions free, 
can be used. If one or more of the hydrogens are replaced by other groups, 
coupling is still possible, provided the coupling positions are left vacant. 
Ortho-hydroxy-diphenyl has the compo